synthesis of small-ring benzannulated triphospha- macrocycles by template...
TRANSCRIPT
Synthesis of small-ring benzannulated triphospha- macrocycles by template methods
Wenjian Zhang
UMI Number: U20219B
All rights reserved
INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a com plete manuscript and there are missing pages, th ese will be noted. Also, if material had to be removed,
a note will indicate the deletion.
Dissertation Publishing
UMI U202193Published by ProQuest LLC 2013. Copyright in the Dissertation held by the Author.
Microform Edition © ProQuest LLC.All rights reserved. This work is protected against
unauthorized copying under Title 17, United States Code.
ProQuest LLC 789 East Eisenhower Parkway
P.O. Box 1346 Ann Arbor, Ml 48106-1346
CardiffU N I V E R S I T Y
PRi F Y S G O L
C aeRDV[§)
BINDING SERVICES Tel+44 (0)29 2087 4949 Fax+44 (0)29 20371921
e-mail [email protected]
Supervisor: Prof. Peter G Edwards
Declaration
Statement 1
This work has not previously been accepted in substance for any
degree and is not being concurrently submitted in candidature for any
degree.
Date ......
Signed V/X...... .. fr. (candidate)
Statement 2
This thesis is the result of my own investigations, except where
otherwise stated. Other sources are acknowledged by footnotes giving
explicit references. A bibliography is appended.
Signed (candidate)
Date ........ ....................................
Statement 3
I hereby give consent for my thesis, if accepted, to be available for
photocopying and for inter-library loan, and for the title and summary to
be made available to outside organisations.
Signed .... .vJN.."................V............. (candidate)
Date ........ ....................................
Chapter 1
Introduction
Chapter 1 Introduction
Foreword
Phosphine ligands, especially tertiaiy phosphines, are one of the most
common and important classes of ligand in organometallic and
coordination chemistry.*1-61 Transition metal phosphine complexes have
broad application in, for example, catalysis, biomedical imaging and
medical therapy. The field of phosphine ligands and their metal
complexes is still growing rapidly due to their academic and industrial
importance as well as new insights which have been developed during
previous decades, particularly on the basis of ligand design for the
catalytic transformation of organic functional groups with high selectivity
and activity. Some of the numerous applications include olefin
metathesis, activation of small molecules such as CO and C-H bond
activation.17141
Phosphine ligands are known to stabilize metal centres over a range of
oxidation states. The most common application of phosphines is as
spectator ligands rather than their direct involvement in reactions as
reagents.1151 One problem of using monophosphines in catalytic
applications is dissociation of the phosphine from the metal centre which
leads to decomposition. This is especially pronounced for the early
transition metals since lability in these systems is commonly high. An
interesting observation is that cyclopentadienyl complexes of the
electropositive transition metals are now well known to form highly
active alkene polymerisation catalysts whereas phosphine complexes of
the same metals are inactive or poorly active. This inactivity appears to
be related to excessive lability as discussed below. The mis-match of soft
P donor and oxophilic hard metal centre contributes to the phosphine
ligand lability under reaction conditions. As multidentate phosphine
ligands have improved ligating ability over monophosphines, macrocycle
1
Chapter 1 Introduction
phosphine ligands should form more stable complexes than their
monodentate analogues. Macrocycle phosphines however, have not been
extensively studied due to synthetic difficulties in accessing such ligands
and hence a lade of suitable examples for study.
Chapter 1 Introduction
1.1 General properties of phosphine ligands
1.1.1 Electronic factors
Two components of the bonding are of importance between phosphine
ligands and metal centres (Figure 1). Phosphine ligands can form a-type
interactions with metals by donating the lone pair of the P atom into an
empty metal orbital with proper symmetry. Phosphines are also known to
act as 7t-acid ligands. The a* orbital of the P-R bonds are believed to
accept the electrons from electron rich metals (back-donation).[15'1?1
Figure 1 Bonding between metals and phosphine ligands
I s /•"'c> M O + O P<a *■ c> M P ^ o
/ /m ) + O p p ► m O p q\ \(a) o bonding between metal and phosphine
OJ)G ^ n . r
dTT
(b) 71 back-donation of electron density from metal to ligand orbital
The strength of a bonding is influenced by the Lewis acidity and basicity
of the metal and ligand respectively. ‘Soft’ P donor phosphines are
generally used to coordinate soft later transition metals. The ability of
phosphine ligands to donate the lone pair can generally be scaled via the
pKa value of the conjugated acid (Eq. 1.1), HPR3+, since there can only
Chapter 1 Introduction
be a a-interaction involved in P-H bonding.[18,191 If the R groups of PR3
are electron-donating, the basicity of phosphines is enhanced (Table 1).
HPR3+ PR3 + H+ (1.1)
[H+l [PR3IK a = — —
Table 1 pKa values of selected phopshines
Phosphine p h 3 PPh3 PPhMe2 PEt3 PBu's
pKa -14 2.73 6.5 8.69 11.40
The change of P-R bond lengths measured by crystallography for
phosphine complexes of the same metal varying in oxidation states, [rj3-
(CgHi3)Fe {P(OMe)3}3]1** (n = 0, 1), confirms the back-donation of
electron density from the metal to the P-R anti-bonding orbitals (i.e. confirms 71-acceptor behaviour).1161 In forming a 71-bond, the strength of
the M-P bond is also enhanced and causes a shortening of the M-P bond
length.*201
The degree of 7c-acidity is also influenced by the nature of the phosphine
substituents. For alkyl groups, 7r-acidity is typically weak; for aryl,
dialkylamino and alkoxy groups, the rc-acidity increases. The 7r-acidity of
PF3 is as great as that of CO due to the strongly electron-withdrawing
property of the F atoms.*151 The order of increasing 71-acid character for a
series of ligands is PMe3 ~ P(NR2)3 < PAr3 < P(OMe)3 < P(0Ar)3 < PC13
< CO ~ PF3.*151 The magnitude of this effect can be quantified by
comparison of CO stretching frequencies in the IR spectrum of a series of
complexes, LNi(CO)3 , L = phosphine ligand (Table 2).[21] The greater
the Ji-acceptor ability of the phosphine ligand, the greater the competition
for electron density from the metal, which reduces the available electron
Chapter 1 Introduction
density for back-donation to the ir* orbital of the CO ligands. This results
in the strengthening of the C-O bond and a higher frequency for v(CO).
Table 2 CO stretching frequencies of some LNi(CO)3 complexes
Phosphine v(C=0) (cm"A) phosphine (cm 1) v(C=0) (cm 1)PC13 2104 PPhMe2 2066P(OPh)3 2086 PMe3 2064P(o-Tol>3 2072 PCy3 2057PPh3 2069
Chapter 1 Introduction
1.1.2 Steric factors
The steric properties of ligands also impose influences on the reactivity of
metal complexes. The coordination sphere of a given metal contains a
limited number of ligands. Bulky phosphine ligands generally stabilize
coordinatively unsaturated metals and are commonly more labile. This
allows small organic substrates to access the vacant site of activated
species.
Steric bulk of PR3 ligands on binding can be measured via Tolman’s cone
angle, the angle of a cone which encompasses the Van der Waals surface
of the phosphine ligand substitutents with the apex at die metal and the R
groups folding back as far as they are able in a space-filling model
(Figure 2) . [221 Hie cone angles of some phosphines are list in Table 3.
Figure 2 Definition of Tolman’s cone angle with PMe3 as an example
Chapter 1 Introduction
Table 3 Tolman’s cone angle for selected phosphines*
Phosphine Cone angle (°) phosphine Cone angle (°)
PMes3 212 PPh3 145
P(o-Tol)3 194 PPr3 132
P B u *3 182 PC13 124
PCy3 170 PMe3 118
* Tolman’s original cone angle is measured from R3PNi(CO)3 complexes
(i.e. where M-P = 2.24 A)
Chapter 1 introduction
12 Macrocycle ligands
Macrocycle ligands may be defined as cyclic organic compounds
containing at least three heteroatom donors within a ring of at least nine
atoms am! with at least two carbon atoms (or dinuclear heteroatoms)
between each adjacent pair of heteroatoms (Figure 3).[23a,bl This is the
definition adopted in this thesis.
Figure 3 examples of N, O, S containing macrocycle
1.1 porphyrin 1.2 [9]aneS3 13 dibenzo-18-crown-6
Macrocycle complexes are widely involved as active centres for
biological processes (eg. Fe complexes of porphyrin ring systems in
electron transport). The synthesis of macrocycle ligands and their
complexes, in order to mimic biological functions, is an important
research topic.f23b] In addition, macrocylic ligands and related complexes
have extensive application in areas of catalysis, clinical diagnosis,
molecular recognition etc.[23] These properties are often based on the
ability of macrocylic ligands to form very stable complexes and also their
selectivity for different metals.[23a,b]
x
Chapter 1 Introduction
1.2.1 Chelate and macrocyclic coordination effects
Multidentate ligands form more stable chelate complexes than related
monodentate ligands {i.e. with otherwise related structure) due to the
chelate effect[23a,bl The chelate effect is primarily entropic in origin. As
illustrated in Eq. 1.2, the substitution of NH3 for en results in an
increased number of species released into solution, whereas the enthalpy
remains approximately unchanged.[23a,b]
[M(NH3)6]2+ + 3en [M(en)3]2+ + 6NH3 (1.2)
4 Species 7 Species
Macrocycle ligands generally bind metal ions into a cavity surrounded by
donor atoms via three or more fused five, six or seven membered chelate
rings. Macrocylic ligands usually form kinetically and
thermodynamically more stable complexes with metal ions than do
acyclic multidentate ligands due to the macrocyclic coordination
effect.p3a,bl The origin of the macrocyclic coordination effect is based not
only on entropy but also on enthalpy.[23a,bl For example, the data in table
4 show the AG for the formation of the cyclic Ni(ii) complex with 1.4 is
more negative than that for the formation of the acyclic complex with 1.5.
Table 4 Thermodynamic data for the formation of Ni(ii) complexes at 298 K[23bl
Low-spin High-spinligand AH/kj TAS/kJ AH/kJ TAS/kJ
mol mol'1 mol'1 mol'11.4 -78.2 49.3 -1 0 0 .8 24.31.5 -66.1 21.7 -80.3 10.9
n iNH HN
^NH HN^uA A
nNH HN
C JN N 'H2 H,
9
Chapter 1 Introduction
1.2.2 Selective Complexation of macrocycle ligands[23a 25]
The property of selective complexation of macrocycle ligands in terms of
stability of the complexes depend on the composition and structure of the
molecule. It is firstly based on the hardness of die donors in the
macrocycle ring. Macrocycle ligands containing hard N or O donors
show different complexation abilities to metal ions compared with
ligands containing soft S donors. The architecture of macrocycle ligands
also plays a role in the selective coordination to metal ions in several
aspects. Firstly, an increased size of the chelate ring usually leads to a
drop in complex stability, five-membered chelate rings are more stable
than six (or more) membered chelate rings. This is due to the greater
difficult in putting the dipoles and charges on donor atoms together and
the increased steric strain as the size of chelate ring increases. Secondly,
the different size of the macrocycle cavity can match different sizes of
metal ions and influences the metal-ligand stability. A good match leads
to a stable complex, mismatch to destabilization. Thirdly, the
conformational flexibility of the macrocycle ring affects the fit of the
metal ion to the ligand and the ability of the ligand to dissociate due to
restricted degrees of freedom in bond rotation. For macrocycles with
saturated aliphatic backbones that can form different conformers by
rotation about single bonds, the ligand may be preorganized for various
metal ion sizes. For macrocycles with more rigid aromatic backbones,
conformational changes in the ligand are relatively restricted so that
movement of the metal centre out of the preferred bonding configuration
is then only possible by major distortions of the coordination sphere.
10
Chapter 1 Introduction
13 Triphosphamacrocycles
Phosphine macrocycles have been investigated for over three decades.
Compared with well-known and extensively studied oxygen, nitrogen and
sulfur containing macrocyles, reports on phosphorus macrocycles are
quite rare. Most published work is still focused on ligand synthesis.11’26-281
Some of the known phosphorus containing macrocycles are shown in
Figure 4.
Figure 4 Some known phosphorus containing macrocycles[1726-291
(a) Triphosphamacrocycles with different ring sizes
9 - membered 12 - membered1.6 1.7 1.8
15 - membered
1.9
CO COR - (CH2)5CH=CH(CH2)5
45 - membered
1.10
(b) Tetraphosphamacrocycles
i.ii 1.12 1.13
1.14 1.15
11
Chapter 1 Introdaction
(c) Macrocycles with mixed N, O, S, P atoms
PhP—P— V i - P —Ph
V
1.16
i ^
OC X )O ' ”
1.19
p------ Li-----I I
1.17
no c x
PhP J PPh
j X Ph
OC, XPhP'[^j'PPh
1.18
P — Ph
V
1.20 1.21
PPh
S S
1.22
PPh
1.23 1.24
\
"~p p-/ I
E.
/ \ / \ PPhyr-O OC \o
O o ypph
/ \ / \ PPh/-N H O
f \NH
*Vnh o /PPh
1.25 1.26 1.27
12
Chapter 1 Introduction
Of the known phosphorus containing macrocycles, triphospha
macrocycles ( P 3 macrocycle) with three syn,syn,syn-P donors are of much
interest due to their electronic analogy to Tj5-Cp" as tridentate 6e' donors
and structure character. P3 macrocycles are facially capping ligands
occupying three coordination sites (as r|5-Cp' ligand) that can impose
important structural influences on the metal centre (Figure 5). A primary
and major influence is that they force the remaining reaction sites into a
mutually cis arrangement. They may also cause shape distortions upon
the coordination sphere with implications for modified reactivity. Due to
their expected ligating ability as soft donor ligands, they may favour
binding the more electronegative metals and those in low oxidation states.
Enhanced stability due to the macrocyclic coordination effect may also
allow access to new classes of complexes of electropositive metals and
also stabilise active intermediates during catalytic reactions etc.
Figure 5 Coordination model of triphosphamacrocycles
* • " r
It is noteworthy that the development of single-site catalysts for olefin
polymerisation is currently dominated by metallocene complexes.1301 The
development of other non-metallocene complexes for catalytic olefin
polymerisation has been of much interest during the last decade.130,311 t j 5 -
Cp" analogous ligands, such as pyridine-diamide,1321 amidodiphosphine,1331
tris(pyrazolyl)borate1341 etc. ligands, have drawn much attention (Figure
6). Triphosphamacrocycles indeed also show intrinsic properties in this
13
Chapter 1 Introduction
application and it is of interest to note their similarities to
cyclopentadienyl ligands as discussed above (see below).
Figure 6 Some tridentate capping ligands analogous to Tj5-CpR as 6 e' donors
pyridine-diamide amidodiphosphine tris(pyrazolyl)borate
1.28 1.29 1.30
1.3.1 Synthetic aspects for the synthesis of triphospha
macrocycles
Despite the expectation that P3 macrocycles will be of value in numerous
applications, studies on P3 macrocycles are somewhat limited due to the
difficulties in synthesis. Two synthetic approaches for the synthesis of
triphosphamacrocycles have been studied namely direct methods and
template-directed methods.
Direct methods:
The phosphine precursors prior to ring closure are combined in dilute
solution in a given ratio to yield the desired phosphine macrocycle. This
method is less effective and has significant drawbacks.
A) The side reaction of polymerisation is hard to avoid even in highly
dilute solution which causes the overall yields to be quite low and
variable.
14
Chapter 1 Introduction
B) Handling the highly volatile, noxious, toxic and air-sensitive
experimentally difficult and also leads to competing decomposition.
C) The stereochemistry o f the phosphine macrocycle can not be
controlled. The product would be obtained as a mixture of various
possible diastereomers. Interconversion of these may be difficult due to
the relatively high inversion energy barrier at P.
The first 11 membered triphosphamacrocycle (1.7) was prepared by Kyba
using a direct method via reaction of a dilithium organophosphide with
halogenophosphine under high dilution conditions (Scheme 1).[3S] The
reaction gave 1.7 as a mixture of stereoisomers that required subsequent
separation. Overall chemical yields were moderate and yields of pure
isomers correspondingly low.
Scheme 1 Synthesis of triphosphamacrocycle by direct method
Template-directed methodsPhosphine precursors are coordinated to a metal centre followed by a
series of intramolecular P-C coupling reactions leading to the desired
phosphine macrocycle. The metal holds the phosphine precursors in close
proximity to facilitate the ring closure reaction effectively and allow the
formation of kinetic products, unlike the thermodynamic products
produced by the direct method. The template-directed method produces
phosphine macrocycles with controlled stereochemistry due to the P
phosphine precursors and intermediates in large volumes of solvent is
High dilution
1.7
15
Chapter 1 Introduction
precursors being preorganised. There are also disadvantages associated
with the template method however.
A) Few suitable metal templates meet the conditions for the synthesis of
phosphine macrocycles are few.
B) Once the phosphine macrocycle has been formed, the macrocycle
complexes are typically kinetically inert and thermodynamically stable
and may be resistant to demetallation.
The first 12-membered triphosphamacrocycle complex (1,5,9-
triphosphacyclododecane, 1.8) was prepared in high yield by Norman
using the yoroMo(CO)3 fragment as a template via the radical-activated
intramolecular hydrophosphination reaction of coordinated
allylphosphine in high yield (Scheme 2).(361
Scheme 2 Template Synthesis of {[12]ane-P3H3}Mo(CO)3
A IB N
1.8
The liberation of free 1,5,9-triphosphacyclododecane (1.31) was achieved
by oxidation of Mo° (d6) to the more labile Mo11 (d4) followed by
hydrolysis with base in our laboratory (Scheme 3).,3?1 The synthesis and
liberation can also be achieved using the alternative fac-Cr(CO)3 or
CpFe+ templates.138 01 A series of derivatives of 1,5,9-triphosphacyclo
dodecane were prepared by alkylation of the secondary phosphines and
related complexes have been investigated. These show interesting
catalytic activity, particularly for olefin polymerization and olefin
metathesis.141,421
16
Chapter 1 Introduction
Scheme 3 Functionalisation and liberation of [12]ane-P3 macrocycles
o c - ci 0 CO
M = M o, Cr
X2/CH2C12
i) 3BuLi ii) 3RX
or alkene/AIBN-► \ \ *
M
oc 1 Nn\ : o co
X'"'° | ^ xX
,OH- OH-
R('T y v
0 'R
OH-
1.31
Br2/CH2C12
17
Chapter 1 inlrodnction
1.3.2 Triphosphamacrocycles with smaller ring sizes
As addressed above, the size of the macrocycle ring leads to selective
complexation of macrocycle ligands. It is well known that [12]ane-N3H3
(1.33) complexes are significantly less stable than are analogous [9]ane-
N 3 H 3 (1.32) complexes (Figure 7).[431
Figure 7 Stability constants of Cu2+ complexes with [9]ane-N3H3 (1.32)
and [12]ane-N3H3 (1.33) ligands
logP = 31.5Cu([9]aiie-N3H3>22+ : ........... w Cu2+ + 2([9]ane-N3H3)
logP = 20.8Cu([ 12]ane-N3H3>22+ w Cu2+ + 2([12]ane-N3H3)
As the chemistry of the 12-membered 1,5,9-triphosphacyclododecane and
related derivatives is developed in more detail, it shows that these ligands
remain flexible. This limits their utility to stabilize f-elements as an
example. Triphosphamacrocycles with ring sizes smaller than 12 atoms,
especially for those with 9-membered rings, should form more robust
complexes due to such macrocycle ligands bind the meter centre with 3
fused 5-membered chelate ring. In this regard, a range of
triphosphamacrocycle complexes with smaller ring sizes have been
prepared using q5-CpRFe+ templates (R = H, alkyl) by the strategy of
intramolecular hydrophosphination or dehydrofluorination.
A) Hydrophosphination
We have previously established the template promoted cyclisation of
alkenyl phosphines on CpRFe+ templates (Scheme 4). Phosphine
precursors were introduced to the iron centre sequentially by replacing
labile ligands (CH3CN in the scheme) which are tram to the CpR ligand,
followed by ring closure reactions. A range of triphosphamacrocycle
H H
N N ^H j^Htacn
[9]ane-N3H3 [12]«ne-N3H3
18
Chapter 1 Introduction
complexes with different ring sizes, including 9 (1.34,[44] 1.381451), 10
(1.35,1461 1.36[471) and 11 (1.37[4T|) membered rings, have been synthesised
via this strategy by varying the phosphine precursors. Aromatic bridges
can also be introduced into the macrocycle backbone (1.38).
Scheme 4 Synthesis of triphosphamacrocycles with small ring sizes by
CpRFe+ templates via hydrophosphination reactions
h2 p n*h2 ^
1.34
9H,
L l l K ,r\» ,t\ y r h \ / 9
hvL=CH,CN
1.35HjP
c Base \ ! /F| ® Z. Fe ® * . Fe ©
R
1.36
vp*rAA / N ^ b L s/VNX
PI.
R
1.37
PHa PHjH"P / p H-VP'Np' ^\ X *> H V. / /*/* ~^Z--- ** Fe ©ii) Base r F w
“RL = C H jC N
R
1.38
19
Chapter 1 Introduction
B) Dehydrofluorination
CpRFe+ templates also support nucleophilic substitution of fluoride on
adjacent aryl groups by coordinated phosphide. This allows the closure of
very rigid 9-membered macrocycles bearing o-phenylene backbone
functions (1.39, 1.40, 1.41)[48,49] as shown in Scheme 5 below.
Alternative donors eg. As, may also be incorporated using this route.1491
Scheme 5 Base promoted intramolecular dehydrofluorination
© © ©PhPAr,
h 2p s' \PH- H,;fpp h 2
Ph4 H2P
h 2p
KOBu
-ToL = CH3CN
©Ar2P PAr2
-►
©H2PPh
e
CH3CN Ar2P ' | x pAr2 CH3CN Ar2P °\ L / \
©
PAr
H2PPh
KOBu
I\PHyPAr2
Ph
KOBu
©
1.40
20
Chapter 1 fatrodncfion
13 3 Template considerations for the synthesis of triphospha-
macrocydes
For the synthesis of triphosphamacrocycles, the choice of suitable
templates is important. Prior to our study of CpRFe+ templates, the
caibonyl complexes of group VI metals with a d6 electron configuration
are the only ones known for the synthesis of 1,5,9-
triphosphacyclododecanes. Other than Kyba’s 11-membered ring systems,
other triphosphorus macrocycles were unknown. For other
triphosphamacrocycle systems with ring sizes smaller than 12 atoms, the
rj5-CpRFe+ templates have proved to be effective. The common feature of
all these templates is that they bear octahedral metal centres with a d6
electron configuration coordinated by inert facially capping ligands (3 x
CO for Cr and Mo, q5-CpR for Fe) and labile ligands (CH3CN, CO or r|6-
C6H3Me3) treats to the capping ligands. These LM (L = inert spectator
ligand) fragments remain intact under the reaction conditions required. It
is also clearly important that the labile ligands can be readily replaced by
suitable phosphine precursors in order to obtain appropriate precursor
phosphine complexes prior to ring closure.
Structural models indicate the P—P non-bonded distances in 1,5,9-
triphosphacyclododecanes are around 3.3 A for the ([ 12]-ane-P3)Cr(CO)3
complex and about 3.5 A for the related Mo complex whereas these
distances are around 2.9-3.0 A in the CpRFe+ complex of 1,4,7-
triphosphacyclononane. The smaller relative atom radius of Fe in
comparison to Cr and Mo (1.26 A for Fe, 1.28 A for Cr, 1.39 A for Mo1501)
results in a closer proximity of the phosphine precursor ligands in the
Fe template complexes, which in turn facilitates the formation of shorter
bridging links between the phosphines and hence smaller macrocycle ring
sizes should be available. The longer Cr-P and Mo-P bond lengths (and
21
Chapter 1 hrtrodnctian
consequently P—P distances) limits the size of the ring system achievable
on the Cr and Mo templates. In addition, for Cr and Mo, the remaining
spectator ligands (3 x CO) are not sterically demanding and leave little
steric influence in ‘pushing’ the phosphines closer together. In the CpRFe
template however, the steric properties of the spectator (CpR) ligand may
be modified to have a significant influence over the trans phosphines (i.e. compressing P-Fe-P angles) aiding closure of smaller ring sizes.
A advantage for the CpRFe+ templates is that the phosphine precursors
can be sequentially introduced into the Fe centre by the use of a range of
cyclopentadienyl derivatives, giving access to a much wide range of
precursor phosphine complexes and incorporation of functionality on the
macrocycle backbones and substituents on the P atoms.
22
Chapter 1 Introduction
1.4 Aim of this thesis
Of all of the template systems investigated to date, the liberation of the
free phosphorus macrocycle has only been achieved with the 1 2 -
membered 1,5,9-triphosphacyclododecane and its tertiary phosphine
derivatives and only readily from Cr° or Mo° carbonyl templates. The
synthesis of triphosphamacrocycles with smaller ring systems has been
achieved by using T]5-CpRFe+ templates. The complexes are however
resistant to liberation of the macrocyle (9aneP3Et3 can only be liberated
from the Fe centre by oxygen transfer oxidation to give the trioxide of
9 aneP3Et3C>3 which is resistant to reduction).*511 The smaller ring systems
are of substantial interest in that they are expected to form very robust
complexes indeed; they remain an important synthetic target.
In this thesis, the application of facially capping ligands to form robust
metal-ligand fragments that will survive the necessary reaction
conditions, and which will limit remaining reaction sites to mutually cis
configurations as alternative templates for the synthesis of
triphosphamacrocycles have been investigated. The aim of this study is
also to allow ready access to the free (uncoordinated) macrocycles.
O f the robust metal-ligand fragments, piano-stool complexes with
octahedral d6 metal centres are promising as clearly demonstrated by the
template-directed synthesis of 1,5,9-triphosphacyclododecane and 1,4,7-
triphosphacyclononane. Examples are, however, few. In our laboratory,
cyclopentadienyl complexes of cobalt, ruthenium and manganese, which
are iso-structural and iso-electronic to the corresponding iron complexes
have been investigated but have not led to new syntheses of macrocycles.
In this thesis, tripodal complexes of trispyrazolylmethane, the
cyclobutadienyl ligand and manganese carbonyl complexes were studied
23
Chapter 1 Introduction___________________________________________________________________
as alternative templates with the aim of developing syntheses of new
triphosphamacrocycles with small ring systems.
24
Chapter 1 hrtrodnctioo
References:
1. C. A. McAuliffe, Comprehensive Coodination Chemistry, Vol. 2, Perganon Pnffl
Oxford, 1987, P989.
2. A. Pidcock, Transition Metal Complexes of Phosphorus, Arsenic and Antimm Ligands, Macmillan, London, 1973, P3.
3. C. A. McAuliffe, W. A Levason, Phosphorus, Arsenic andStibine Complexes the transition Elements, Elsevier, Amsterdam, 1979.
4. T. Appleby, J. D. Woollins, Coord. Chem. Rev. 235, 2002, 121-140.
5. H. A Mayer, Chem. Rev. 1994, 94,1239-1272.
6. P. Braunstein; N. Boag, Angew. Chem. Int. Ed. 2001,40(13), 2427-2433.
7. T. Hayashi, Ace. Chem. Res. 2000,33(6), 354-362.
8. E. Milko, D. Milstein, Chem. Rev. 2003,103(5), 1759-1792.
9. H. H. Karsch; V. Graf; M. Reisky, Phosphorus, Sulfur and Silicon and to Related Elements 1999,144-146, 553-556.
10. U. Segerer, R. Felsbeig; S. Blaurock; G. A. Hadi; E. Hey-Hawkins, Phosphorn
Sulfur and Silicon and the Related Elements 1999,144-146, 477-480.
11. H. Grutzmacher; C. Meyer, S. Boulmaaz; H. Schonberg; S. Deblon; J. Liedtke;S
Loss; M. Worle, Phosphorus, Sulfur and Silicon and the Related Elements 1999,
144-146,465-468.
12. K. V. L. Crepy; T. Imamoto, Topics in Current Chemistry 2003, 229 (New
Aspects in Phosphorus Chemistry III), 1-40.
13. T. Appleby; J. D. Woollins, Coord Chem. Rev. 2002,235(1-2), 121-140.
14. D. Mlngos; P. Michael, Modem Coordination Chemistry 2002, 69-78.
15. R. H. Crabtree, The Organometallic Chemistry of the Transition metals, third
edition, John Wiley & Sons, New York, 91-95,2001.
16. A. G. Orpen, Chem. Commun. 1310,1985
17. G. Pacchioni; P. S. Bagus, Inorg. Chem. 31, 4391,1992
18. C. A. Streuli, Anal. Chem., 1960, 82, 5791
19. T. Allman; R. G. Goel, Can. J. Chem., 1982,60, 716
20. M. J. Wovkulich; J. L. Atwood, L. Canada, J. D. Atwood, Organometallics, 1985,
4,867
21. C. A. Tolman, J. Amer. Chem. Soc. 1970, 92,2953
22. C. A Tolman, Chem. Rev., 1977, 77, 313
25
Chapter 1 Introduction
23. (a) L. F Lindoy. The Chemistry of Macrocyclic Ligand Complexes 1992, 269 pp
Science, (b) E. C. Constable, Coordination Chemistry ofMarocyclic Compounds, Oxford. 1999. (c) D. Salvemini; D. P. Riley, Cellular and Molecular Life Sciences 2000, 57(11), 1489-1492. (d) H. G. Jr. Richey, Comprehensive Supramolecular Chemistry 1996, 1,755-775.
24. P, Comba; W. Schiek, Coord Chem. Rev. 238-239, 2003, 21-29
25. R. D. Hancock; A. E. Martell, Chem. Rev. 1989, 89,1875-1914
26. W. Levason; G. Reid, Comprehensive Coordination. Chemistry LI, 2004,1,475-
484.
27. O. Stelzer; K.-P. Langhans, The Chemistry of Organophosphorus Compounds, F.
R. Hartley, Eds., Wiley, New York, 1990, vol 1. Ch8, 192-254
28. A. Caminade; J. P. Majoral, Chem. Rev. 1994,94(5), 1183-213
29. S. Eldci; M. Nieger, R. Glaum; E. Niecke, Angew. Chem. Int. Ed. 2003, 42(4),
435-438
30. (a) W. Wang; L. Wang, J. Poly. Mater. 2003,20(1), 1-8. (b) G. G. Hlatky,
Coord Chem. Rev. 1999,181,243-296.
31. G. J. P. Britovsek; V. C. Gibson; D. F. Wass, Angew. Chem. Lnt. Ed. 1999, 38(4),
428-447.
32. F. Guerin; D. H. McConville; J. J. Vittal, Organometallics 1996, 15, 5586-5590
33. M. D. Fryzuk; G. Giesbrecht; S. J. Rettig, Organometallics 1996, 15, 3329-3336
34. S. Scheuer, J. Fisher, J. Kress, Organometallics 1995,14, 2627-2629
35. E. P. Kyba; A. M. John; S. B. Brown; C. W. Hudson; M. J. McPhaul; A . Harding;
K. Larsen; S. Niedzwiecki; R. E. Davis, J. Am. Chem. Soc. 1980,102(1), 139-47.
36. B. N. Diet; R. C. Haltiwanger; A. D. Norman, J. Am. Chem. Soc. 1982, 104(17),
4700-1.
37. S. J. Coles; P. G. Edwards; J. S. Fleming; M. B. Hursthouse; S. S. Liyanage,
Chem. Common. 1996, (3), 293-4.
38. P. G. Edwards; J. S. Fleming; S. S. Liyanage, Lnorg. Chem. 1996, 35(16), 4563-
4568.
39. P. G. Edwards; J. S. Fleming; S. S. Liyanage, Dalton Trans., 1997, (2), 193-197.
40. A. J. Price; P. G. Edwards, Chem. Commun. 2000 (11), 899-900.
41. R. J. Baker, P. G. Edwards, Dalton Trans. 2002, (15), 2960-2965
42. R. J. Baker; P. G. Edwards; J. Gracia-Mora; F. Ingold; K. M. A. Malik, Dalton Trans. 2002, (21), 3985-3992.
26
Chapter 1 Introduction
43. R. M Smoth; A. E. Martell, Critical Stability Constants, Vols. 5 and 6, Plenum,
1989
44. P. G. Edwards; P. D. Newman; K. M. A. Malik, Angew. Chem, Int. Ed. 2000,
39(16), 2922-2924.
45. P. G. Edwards; M. L. Whatton; R. Haigh, Organometallics 2000,19(14), 2652-
2654.
46. P. G. Edwards; P. D. Newman; D. E. Hibbs, Angew. Chem. Int. Ed. 2000,39(15),
2722-2724.
47. R. Haigh, Doctoral Thesis, 2002, Cardiff University
48. T. Albers, Doctoral Thesis, 2001, Cardiff University
49. J. Johnstone, Doctoral Thesis, 2004, Cardiff University
50. J. G. Speight, Lange \s Handbook of Chemistry 16th Ed. McGraw-Hill, New York
2005.
51. P. G. Edwards, P. D. Newman, Cardiff University; Unpublished results.
27
Chapter 2
Tripodal Tris-py razoly lmethane Iron
and Cobalt Complexes
Chapter 7. Tripodal TrispvrazoMmethane Iron and Cobalt Complexes
Abstract
A series of cationic trispyrazolylmethane complexes of the general form
[TmRM(CH3CN)3]2+ (Tm = tris(pyrazolyl)methane, 2.1, R = 3 ,5 -Me2, M
= Fe; 2J2, R = 3-Ph, M = Fe; 23, R = 3 ,5 -Me2, M = Co; 2.4, R = 3-Ph, M
= Co) with ‘piano-stool’ structure were prepared by the reaction of the N3
tripodal ligands (TmR) with [(CH3CN)6M](BF4)2 in a stoichiometric 1:1
ratio. Magnetic susceptibility measurements indicate that all four
complexes with BF4 counter anions are paramagnetic, high-spin systems
in the solid state with geff of 5.1 (2.1, S = 2), 5.1 (2.2, S = 2), 4.6 (2.3, S =
3/2) and 4.5 (2.4, S = 3/2) respectively. Comparisons of bond lengths
from the metal centre to the TmR nitrogen donors, and from the metal
centre to the acetonitrile nitrogen donors indicate that the neutral tripodal
ligands appear to be more weakly coordinated to the metal centre than are
the acetonitrile ligands. Reactions of these tripodal complexes with
bidentate phosphine ligands, such as 1 ,2 -diphosphinoethane or 1 ,2 -
bis(diaUylphosphino)ethane leads to displacement of the tripodal ligand,
or to the formation of more thermally stable bis-ligand complexes ML2
{L = tris(3,5-dimethyl-l -pyrazoyl)methane}.
28
Chapter 2 Tripodal Trispvrazolvifnethane Iron and Cobalt Complexes
2.1 Introduction
We are interested in octahedral transition metal complexes stabilised by
tripodal capping ligands, and which also contain labile monodentate
ligands in a ‘piano-stooP type of structure for the following reasons.
These complexes may be potential templates for the synthesis of
triphosphamacrocycles by replacing the labile monodentate ligands with
phosphine presursors which may support intramolecular P-C bond
formation by a range o f suitable reactions as we have demonstrated for
die synthesis of triphosphamacrocycles using the Tj5-CpRFeL3+ (L =
CH3CN) class of complexes as templates (Chapter 1).
Current templates for the formation of triphosphorus macrocycles are
either demetallated with difficulty, and/or with decomposition or not at
all and templates stabilised by alternative tripodal capping ligand may be
more amenable to liberation of the phosphorus macrocycle.
The tripodal nitrogen chelate ligands, tris(pyrazolyl)borate (Tp) and
tris(pyrazolyl)methane (Tm) have been extensively investigated in
inorganic, bioinorganic and organometallic chemistry, partly due to then-
close analogy to cyclopentadienyl ligands.11"21 All these ligands are 6
electron donors (Figure 1) and occupy three coordination sites as capping
ligands when binding a metal. These ligands form a variety of metal
complexes that have counterparts in the classic cyclopentadienyl-iron
systems.121 The Tm or Tp ligands commonly form bis-ligand structures
for octahedral metal ions in which the two tripodal ligands interdigitate,
since the donors are held in a planar fashion roughly, in plane with the
metal-donor bond. The parent tris(pyrazolyl)borate or
tris(pyrazolyl)methane ligands can be modified by introducing
substituents such as -CH3, -C3H7, -C4H9 or -C6H5 in the 3- or 3,5-
29
Chapter 2 Tripodal Trispvrazolylmethane Iron and Cobalt Complexes
positions of the pyrazole ring resulting in variations of steric bulk. These
derivatives with controlled coordination behaviour may prevent
formation of bis-ligand ML2 complexes (L = Tp or Tm ligand).[2] The
known tripodal complexes, [TpRM(NCMe)3]+ (Tp = hydrotrispyrazolyl-
borate; M = Ni, Co), which are very labile in the solution absence of
CH3CN, were prepared by chloride abstraction from the corresponding
chloro complexes, TpRM-Cl.[3] The related reported tripodal complex of
{Fe[(3,5-Me2pz)3CH](H20)3}2+ shows less solubility in organic
solvents.[4) This chapter will introduce the work of the direct and
stoichiometric synthesis of piano-stool Fe/Co complexes of substituted
tris-pyrazolylmethanes, their structural characterisation, magnetic
properties and preliminary reactions with bis-phosphines.
H H
B * C
r y >N
I\ ;
M M
IXFigure 1 Analogy between Cp, tris(pyrazolyl)borate and
tris(pyrazolyl)methane
30
Chapter 2 Tripodal Trispvrazolylmethane Iron and Cobalt Complexes
2.2 Synthesis o f trip o dal complexes
The four tripodal complexes of [TmRM(CH3CN)3](BF4)2 (Tm =
tris(pyrazolyl)methane; R=3-Ph, 3,5-Me2; M = Fe, Co) were directly
[(CH3CN)6M](BF4 )2 in acetonitrile solvent (Scheme 1). The products are
conveniently crystallised directly from the reaction mixture and are
isolated as colourless (Fe) or orange/red (Co) crystals. IR spectra of these
complexes confirm the existence of three terminal CH3CN ligands. The
compounds are all air- and moisture- sensitive as the solids or in solution.
Scheme 1 Preparation of tripodal trispyrazolylmethane iron and cobalt
complexes, [(TMR)M(CH3CN)3]2+
n c c h 3
The thermally stable bis-ligand complexes, [ML2]2+ (M = Fe, Co; L =
hindrance introduced by substituents in the 3, 5 positions of the pyrazole
ring. The ratio of the two reagents in the reaction is also very important in
order to limit formation of the bis-ligand complexes. In the mass spectra
of the iron and cobalt complexes 2.1 to 2.4, the highest ion observed
corresponds to the [TmRM]2+ fragment (m/z = 354, 498, 357, 501 amu for
2.1 to 2.4 respectively) except when the compound is admitted as a
methanol solution in which case complex ions corresponding to the
formula, [TmRM(CH3CN)2(CH3OH)]2+ were observed (m/z = 468, 612,
471, 615 amu for 2.1 to 2.4 respectively) in all cases. Crystalline samples
prepared by reactions of tripodal TmR ligands with one equivalent of
H
M(CH3CN)62+ + TmR CH3CN R,
2+
2.2 M = Fe Rj = -Ph R2 = H
c h 3c n NCCH3 2.4 M - Co R, = -Ph R2 = H
TmR), were avoided due to coordination control through the steric
31
Chapter 2 Tripodal Trispvrazolvlmethane Iron and Cobalt Complexes
of compounds 2 .1 and 2 .2 are colourless, which is consistent with a
paramagnetic high-spin (HS) state for iron complexes at room
temperature.151 In the NMR spectra of these four tripodal complexes in
solution, only broad non-diagnostic resonances are observed due to them
being paramagnetic.
Thermal analyses (TGA) indicated three decomposition steps for
complexes 2.1*2BF4 and 2.3*2BF4 corresponding to loss of coordinated
acetonitrile ligands between 80-190 °C with weight losses of 18.9% and
18.8% (calculated amount) respectively followed by loss of one molecule
of BF3 between 190 to 240 °C with weight losses of 10.4% and 10.3%
(calculated amount) respectively, then loss of tris(3,5-dimethyl-
pyrazoyl)methane up to 400 °C, with weight losses of 45.7% and 45.5%
(calculated amount) respectively (Figures 2, a and c).
The thermal decomposition of complex 2.2-2BF4 was complicated
(Figure 2, b). The loss of acetonitrile donors between 80 to 190 °C
(15.5% weight, calculated) and a BF3 molecule between 190 to 225 °C
(8.5% weight, calculated amount) were observed. The tris(3-
phenylpyrazoyl)methane ligand probably begins to be decomposed at
around 300 °C.
For complex 2.4-2BF4-CH3CN, multi decomposition steps were involved
upon heating (Figure 2, d). The steps of losing the co-crystallised
acetonitrile and the three coordinated acetonitrile ligands between 60-190
°C (19.5% for the four CH3CN, calculated amount) were not well-
separated. The loss of BF3 (8.0% weight, calculated amount) was
observed between 190 to 250 °C. From 250 to 550 °C, an observed total
weight loss of 41% was assigned to the decomposition of the tris(3-
phenylpyrazoyl)methane ligand (52.6%, calculated amount).
32
Chapter 2 Tripodal Trispvrazolvlmethane Iron and Cobalt Complexes
Figure 3 TGA diagrams o f complexes 2.1-2BF4 (a), 2.2-2BF4 (b),
2.3-2BF4 (c), and 2.4-2BF4 CH3CN (d)
i s
10D
BO
I "jfi
30
SOD
(a)
i s
Cll.131
£I
OBI Mil)
10
EDO
(b)
33
Chapter 2 Tripodal Trispyrazolvlmethane Iron and Cobalt Complexes
I S
I S
EDO
(C)
1®
BO
70too
(d)
Magnetic susceptibility studies also confirm that all four tripodal
complexes with BF4‘ counter anions are high-spin paramagnetic
complexes with p ^ o f 5.1 (2.1, S = 2), 5.1 (2.2, S = 2), 4.6 (2.3, S = 3/2)
and 4.5 (2.4, S = 3/2) respectively in the solid state (Figure 3) at room
temperature. The bis-ligand Fe(II)/Co(II) complexes, (TmR)2M(BF4)2 (M=
34
Chapter 2 Tripodal Trispvrazolvlmethane Iron and Cobalt Complexes
Fe, Co; R = 3,5-Me2), are reported to exhibit spin crossover properties in
the solid state.16,71 For the iron complexes of 2.1-2BF4 and 2.2-2BF4, no
spin crossovers are observed over a temperature range from 4K to 300K.
The magnetic moment of complex 2.3-2BF4 gradually increased from 4.2
to 4.6 as the temperature increased from 4K to 100K whereas the
corresponding value of complex 2.4-2BF4 CH3CN increased from 3.4 to
4.4 as the temperature increased from 4K to 150K which might suggest
very weak spin exchange for the two cobalt complexes at low
temperature.
Figure 3 Magnetic susceptibility diagrams of complexes 2.1-2BF4 (a),
2.2-2BF4 (b), 2.3-2BF4 (c) and 2.3-2BF4 CH3CN (d)
0.300.25
0.250.20
0.20VE 0.133I 0.10
0 0 5
0.15
2 0 10
'ft 0 05
0.000.00300100 200100 200 300
T «
(a)
3 i
T «
(b)
■ft 0 .1 0
O & t*0 0<K>©OGvtf'<k e- C
(c)
susceptlbllty
The structures of complexes 2.1 to 2.4 were characterized by X-ray
diffraction studies as well. Their crystal structures are illustrated in
Figures 4, 5, 6 and 7 respectively. Table 1 gives selected structural
parameters for the four tripodal complexes.
35
Chapter 2 Tripodal Trispvrazolvlmethane Iron and Cobalt Complexes
Figure 4 ORTEP plot (30% probability level) of 2.1 (H atoms omitted)
C7C3
C5
C6N2Fe1 C8
C2
Selected bond lengths (A): Fe(l)-N(2) 2.165(3) Fe(l)-N (l) 2.165(3) Selected bond angles (°): N(2)-Fe(l)-N(2)#l 84.88(11) N(2)-Fe(l)-N(l) 92.11(12) N(2)#l-Fe(l)-N(l) 92.83(12) N(2)#2-Fe(l)-N(l) 176.36(12) N(l)-Fe(l)-N(l)#2 90.08(13) (#1 -y+1, x-y, z; #2 -x+y+1, -x+1, z)
36
Chapter 2 Tripodal Trispvrazolvlmethane Iron and Cobalt Complexes
Figure 5 ORTEP plot (30% probability level) of 2.2 (H atoms omitted)
Selected bond lengths (A): Fel-N l 2.199(2) Fel-N3 2.196(2) Fel-N5 2.205(2) Fel-N7 2.131(3) Fel-N 8 2.156(3) Fel-N9 2.165(3) Selected bond angles (°): N(7>Fe(l)-N(8) 86.37(10) N(7>Fe(l>N(9) 87.64(10) N(8)-Fe(l)-N(9) 90.29(10) N(7)-Fe(l>N(3) 101.57(9) N(8>Fe(l)-N(3) 89.59(10) N(9>Fe(l>N(3) 170.76(9) N(7)-Fe(l>N(l) 172.21(10) N(8 )- Fe(l)-N (l) 98.78(10) N(9>Fe(l>N(l) 86.50(9) N(3>Fe(l)-N(l) 84.40(9) N(7>Fe(l)-N(5) 89.86(9) N(8)-Fe(l)-N(5) 171.10(10) N(9)-Fe(l)-N(5) 97.61(9) N(3>Fe(l>N(5) 83.27(9) N(l)-Fe(l)-N(5 ) 85.84(9) C(1>N(1)- N(2) 104.6(2)
37
Chapter 2 Tripodal Trispvrazolvlmethane Iron and Cobalt Complexes
Figure 6 ORTEP plot (30% probability level) of 2.3 (H atoms omitted)
Selected bond lengths (A): C o(l>N (l) 2.122(5) Co(l>N(2 ) 2.125(4) Selected bond angles (°): N (l>C o(l)-N (l)#l 89.15(18) N (l)-Co(l> N(2)#2 92.42(17) N(l)#l-Co(l)-N(2)#2 92.96(17) N(l)#2-Co(l>N(2)#2 177.39(16) N(2)#2-Co(l)-N(2)#l 85.42(17) (#1 -y+1, x-y, z; #2 -x+y+1, -x+1 , z)
38
Chapter 7. Tripodal Trispvrazolylmethane Iron and Cobalt Complexes
Figure 7 ORTEP plot (30% probability level) of 2.4 (H atoms omitted)
C16
C25
C26
C27 C28
Selected bond lengths (A): Col-N l 2.169(2) Col-N3 2.167(2) Col-N5 2.172(2) Col-N7 2.091(2) C0 I-N8 2.117(2) Col-N3 2.129(2) Selected bond angles (°): N(7>Co(l>N(8) 85.68(9) N(7>Co(l>N(9) 86.64(8) N(8)-Co(l>N(9) 89.25(8) N(7>Co(l>N(3) 101.46(8) N(8>Co(l>N(3) 90.14(8) N(9>Co(l>N(3) 171.81(8) N(7>Co(l>N(l) 172.34(8) N(8 )- C o(l>N (l) 98.56(8) N(9>Co(l>N(l) 87.04(8) N(3)-Co(l)-N(l) 84.98(8) N(7)-Co(l)-N(5) 90.42(8) N(8>Co(l>N(5) 171.84(8) N(9>Co(l>N(5) 97.69(8) N(3)-Co(l)-N(5) 83.60(8) N(l)-Co(l)-N(5) 86.13(8)
39
Chapter 2 Tripodal Trispvrazolvlinediane Iron and Cobalt Complexes
The ideal molecular symmetry for these complexes is C3V with a 3-fold
axis (2.1 and 23) or pseudo 3-fold axis (2.2 and 2.4) which passes
through the metal centre and the apical C atom which links the three
pyrazole rings. All four tripodal complexes have a similar structure
bearing a distorted octahedral metal centre which is coordinated by tris-
pyrazolylmethane ligand in tridentate facially capping model and three
acetonitrile supporting ligands trans to the TmR N donors. The metal
centres are sandwiched by two roughly parallel nitrogen planes, one
defined by the TmR nitrogen donors and the other by the acetonitrile
nitrogen donors. The average Fe-N bond lengths in 2.1 {2.165(3) A} and
2.2 (2.175(3) A} are consistent with the high-spin configuration of the Fe
centre in both complexes at 150K. Average bond lengths of ca 1 .97 A are
typically observed in low-spin (LS) octahedral FeN6 complexes.15,71 The
TmR ligands chelate the metal centre via three fused 6 -membered rings
which restricts the average M-N-N angles to 84.88(11)°, 84.50(9)°,
85.42(17)° and 84.90(8)° for complexes 2.1 to 2.4 respectively.
The Fe-N bond length between Fe and the TmR N donors in 2.1 {2.165(3)
A} is shorter than that in 2.2 {aver. 2 .2 0 0 (2 ) A}, and the corresponding
Co-N bond lengths in 2.3 {2.125(3) A} are also shorter than that in 2.4
{aver. 2.169(2) A}. These M-N bond length changes are consistent with
the variation in the steric bulk of the Tm ligand as a function of the
substituents (R = phenyl, methyl) on the pryazole ring. The Co-N bond
lengths between Co and TmR N donors in 2.3 and 2.4 are both longer than
those {2.150(4) A} in the known [TpRCo(CH3CN)3]+ (Tp =
hydrotrispyrazolylborate, R = 3-Ph, 5-Me) . [31 This implies that the anionic
TpR ligand coordinates more tightly to die metal centre than does the
neutral TmR ligand in complexes 2.3 and 2.4.
40
Chapter 2 Tripndal Trispvrazolvlmethane Iron and Cobah Complexes
In 2.1 and 2.3, the M-N(TmR) value is almost the same as that of the M-
NCCH3 value. In 2.2 and 2.4, die M-N(TmR) values are both much larger
than the M-N(CH3CN) values. The distances of the metal to the N3 plane
centroid (constituted by TmR nitrogen donors) for these complexes are
also longer than the distances of the metal to the N3 plane centroid
constituted by the acetonitrile nitrogen donors. These values imply that
the neutral tripodal ligand coordinates relatively weakly to the metal in
comparison to the acetonitrile ligands. This may be due to the steric bulk
of the 3-substituent on the pyrazole ring.
Table 1 Selected structural parameters for complexes 2.1,2.2,2.3 and 2.4
Complex 2.1 2J2 23 2.4
M-N(TmR)(A)2.165(3) 2.199(2)
2.196(2) 2.205(2)
aver. 2.200
2.125(3) 2.169(2) 2.167(2) 2.172(2)
aver. 2.169
M-N(CH3CN)(A)2.165(3) 2.156(3)
2.131(3) 2.165(3)
aver. 2.150
2.122(4) 2.091(2) 2.117(2) 2.129(2)
aver. 2.112
Aver. M-N (A) 2.165(3) 2.175(3) 2.123(4) 2.140(2)
N-M-N(TmR)(°)84.88(11) 84.40(9)
83.27(9) 85.84(9)
aver. 84.50
85.41(14) 84.98(8) 83.60(8) 86.13(5)
aver. 84.90Distance of M to N3 plane of TmR(A)
1.356(3) 1.386(2) 1.322(3) 1.359(2)
Distance of M to N3 plane of CH3CN (A)
1.248(3) 1.282(3) 1.243(4) 1.277(2)
41
Chapter 2 Tripodal TrispvrazoMmetiiane Iron and Cobalt Complexes
2 3 Reactions of tripodal complexes with phosphine reagents
As a potential template for the synthesis of triphosphamacrocycles, the
tripodal metal-ligand fragments need to be stable when treated with
phosphine ligands whereas the labile monodentate ligands trans to the
tripodal ligand within the coordination sphere, need to be replaced
sequentially by phosphine ligands. Consequently the tripodal complexes
2.1 to 2.4 were reacted with 1 equivalent of the bidentate phosphine
ligands, 1,2-diphosphinoethane (dpe) and 1,2-bis(diallylphosphino)-
ethane (tape) (Scheme 2) respectively.
Scheme 2 Reactions of complex 2.1 to 2.4 with bidentate phosphine
ligands
^ [(TmR)2M]2 for 2.1 and 2.3, M = Fe, Co
[TmRM(CH3CN>3f Free TmR for 2.2 and 2.4
[(tape^FeCCHsCN),]2* (2.5) for 2.2
Reactions with the diphosphines did not produce the expected complexes,
but resulted in displacement of the tripodal TmR ligands. It is likely that
the generally weaker binding between the TmR ligands and the metal
centre (in part due to the steric bulk of the 3-substituent on the pyrazole
ring) leads to lability and the displacement of the TmR ligands from the
metal centre when complexes 2.1 to 2.4 were treated with dpe in solution.
The tris(pryazolyl)methane ligands behave as bidentate ligands in
complexes (M(PMe3)2(COXCOMe)[q2-(3,5-Me2pz)3CH]}+ (M = Fe, Ru;
pz = pyrazolyl ring) may confirms the labile coordination model of such
42
Chapter 2 Tripodal TrispvrazoMmethane Iron and Cobah Complexes
ligands.1*1 In the case of complexes 2.2 and 2.4, the free tripodal TmR
ligand was isolated as white powders and identified by NMR
spectroscopy and melting point determination. For complexes 2.1 and 2.3
with the less stoically hindered (3,5-dimethyl-pyrazoyl)methane ligand, n A | _
the bis-tripodal complexes [(Tm ^M] (M = Fe, Co) were isolated in
low yield (confirmed crystallographically).
The reaction of complex 2.2 with l,2-bis(diallylphosphino)ethane (tape)
gives complex 2.5 which has been characterized by X-ray diffraction.
Figure 8 provides the ORTEP diagram of complex 2.5 with selected
bond lengths and bond angles. As the structure shows, complex 2.5 does
not include the tripodal ligands. The iron centre, with pseudo-octahedral
geometry is located on a crystallographic inversion centre coordinated by
two bidentate tape ligands in the equatorial plane and two acetonitrile
ligands in axial positions.
43
Chapter 2 Tripodal Trispvrazolvlmethane Iron and Cohalt Complexes
Figure 8 ORTEP plot (30% probability level) of 2.5 (H atoms omitted)
C7C6
C10C5
C9C8
C3C4Fe1
P2C15
C16C11
Selected bond lengths (A): Fel-N l 1.905(3) Fel-Pl 2.3032(9) Fel-P2 2.2911(9) Selected bond angles (°): N (l>Fe(l)-N (l)#l 180.00 N (l> Fe(l)-P(2)#l 88.72(9) N(l>Fe(l)-P(2) 91.28(9) N (l>Fe(l)-P(l) 89.33(9) N (l)-Fe(l)-P(l)#l 90.67(9) P(2)-Fe(l)-P(l) 84.03(3) P(2>Fe(l)-P(l)#l 95.97(3) P(2)#l-Fe(l)-P(2) 180.00 P(l)-Fe(l)-P(l)#l 180.00 (#1 -x, -y, -z)
44
Chapter 2 Tripodal Trispvrazolvltnethane Iron and Cobalt Complexes
2.4 Conclusions
Novel tripodal trispyrazolylmethane complexes of Fe(II)/Co(II) with
piano-stool structures supported by acetonitrile ligands have been
prepared and characterised. The preliminary investigation of substitution
reactions of these complexes with phosphine ligands showed that the LM
(L = TmR) fragment is not resistant enough to displacement, to limit the
remaining reaction to sites to be trans to the tripodal ligand. Reactions
with bis-phosphines lead instead however to substitution of the tripodal
ligand. This results in the formation of the more thermally stable bis-
ligand ML2 complex. The removal of the tripodal ligand from the metal
centre illustrates the flexibility of the bonding between the tripodal
ligand and the metal in these complexes. Thus these tripodal metal-ligand
fragments are not suitable as templates for the synthesis of P 3 marocycles.
45
Chapter 2 Tripodal Trispvrazolvl methane Iron and Cohalt Complexes
2.5 Experimental Section
General experimental information: All reactions were performed under
a nitrogen atmosphere with standard Schlenk glassware, vacuum or
Glovebox techniques unless otherwise noted. The solvents were dried and
degassed by refluxing over standard (hying agents and distilled
immediately prior to use. Infrared spectra were recorded on a Perkin
Elmer 1600 spectrometer using KBr pellets for solid samples. Mass
spectra were carried out on a VG Platform II Fisons mass spectrometer.
Magnetic susceptibility measurements were performed with a Quantum
Design MPMS2 SQUID magnetometer in collaboration with Prof.
Andrew Harrison in Edinburgh. TGA measurements were performed on a
SDT Q600 thermal analyzer. Elemental analyses were performed by
Warwick Analytical Service. X-ray diffraction data collections were
carried out on a Bruker Kappa CCD diffractometer at 150(2) K with Mo
Ka irradiation (graphite monochromator). Empirical absorption
corrections were performed using equivalent reflections. For the solution
and refinement of the structures, the program package SHELXL 97 was
employed.t91 H atoms were placed into calculated positions and included
in the last cycles of refinement. Crystal structure and refinement data are
collected in Appendix A and supplementary CD.
Materials: Tris(3,5-dimethyl-pyrazoyl)methane,1101 tris(3-
phenylpyrazoyl)methane,1101 [M(CH3CN)6](BF4)2 (M = Fe, Co),1111 1,2-
diphosphinoethane(dpe),[12] 1,2-bis(diallylphosphino)ethane(tape)[13] were
prepared according to literature methods. All other chemicals were
obtained from commercial sources and, where appropriate, dried over
molecular sieves and degassed by repeated freeze-thaw degassing.
46
Chapter 2 Tripodal Trispvrazolvlfnethane Iron and Cobalt Complexes
[rm RFe(CHjCN)3](BF4)2 (2.1 -2BF4, R = 3,5-Me2)
0.20 g (0.5 mmol) o f tris(3,5-dimethyl-pyrazolyl)methane in 10 ml
CH2CI2 was gradually added to a solution of 0.22 g (0.50 mmol)
[Fe(CH3CN)b](BF4)2 in 20 ml CH3CN with stirring at room temperature.
The solvents were evaporated and 20 ml CH3CN added to dissolve the
solid residue. The solution was filtered and concentrated. Colourless
crystals o f 2.1*2BF4 were obtained by vapour diffusion of diethyl ether
into the filtrate. M.S.(APCI): 354 {[TmRFe]2+, in CH3CN solution}, 468
{[TmRFe(CH3CN)2(CH3 0 H)]2+, in CH3OH solution}. IR (KBr): 2313
( v c n ) , 2283 ( v c n ) . Anal. Calcd for C22H3iN9B2F8Fei (651.2 gmol"1) C,
40.59; H, 4.80; N, 19.36. Found: C, 39.68; H, 4.67; N, 18.85.
[(TmR)Fc(CH3CN)3](BF4)2 (2.2-2BF4, R = 3-Ph)
0 .2 0 g (0.5 mmol) of tris(3-phenylpyrazolyl)methane in 1 0 ml CH2CI2
was gradually added to a solution of 0.22 g (0.50 mmol)
[Fe(CH3CN)6j(BF4)2 in 10ml CH3CN with stirring at room temperature.
After half an hour, the volatiles were removed in vacuo and 10 ml
CH3CN added to dissolve the solid residue. The solution was filtered and
concentrated by evaporating some of the solvent under vacuum.
Colourless crystals of 2.2-2BF4 were obtained by vapour diffusion of
diethyl ether into the filtrate. M.S.(APCI): 498 {[TmRCo]2+, in CH3CN},
612 {[TmRCo(CH3CN)2(CH3 0 H)]2+, in CH3OH}. IR (KBr): 2310 (vCN),
2282 ( v c n ) . Anal. Calcd for Cj+HjiNsBjFgFe, (795.2 gmol'1): C,
51.31; H, 3.93; N, 15.85 Found: C, 50.88; H, 3.85; N, 15.64.
[(TmR)Co(CH3CN)3|(BF4)2 (2J-2BF4, R = 3,5-Mej)
0.20 g (0.5 mmol) of tris(3,5-dimethylpyrazolyl)methane in 1 0 ml CH2CI2
was gradually added to a pink solution of 0.22 g (0.50 mmol)
[ C o ( C H 3 C N ) 6 ] ( B F 4 ) 2 in 20 ml C H 3 C N with stirring at room temperature.
47
Chapter 2 Tripodal Trispvrazolvl methane Tron and Cobalt Complexes
Tthe solvents were removed and 20 ml CH3CN added to dissolve the pink
solid residue. The solution was filtered and concentrated. Yellow
crystalline 23-2BF4 were obtained by vapour diffusion of diethyl ether
into the filtrate. M.S.(APCI): 357 {[TmRFe]2+, in CHjCN}, 471
{[TmRFe(CH3CN)2(CH3 0 H)]2+, in CHjOH}. IR (KBr): 2316 ( v c n ) , 2287
( v c n ) . Anal. Calcd for C22H31N9B2F8C01 (654.2 gmol'1) C, 40.40; H, 4.78;
N, 19.27. Found: C, 39.36; H, 4.84; N, 18.54.
[(TmR)Co(CH3CN)3l(BF4)2 (2.4-2BF4, R = 3-Ph)
The complex was prepared in an analogous fashion to 2.3-2BF4. 0.20 g
(0.5 mmol) of tris(3-phenylpyrazolyl)methane in 10 ml CH2CI2 was
gradually added to a pink solution of 0.22 g (0.50 mmol)
[Co(CH3CN)6](BF4)2 in 20 ml CH3CN with stirring at room temperature.
The solvent were removed and 20 ml CH3CN added to dissolve the pink
solid residue. The solution was filtered and concentrated. Red-orange
crystals o f 2.4-2BF4*CH3CN were obtained by vapour diffusion of diethyl
ether into the filtrate. M.S.(APCI): 501 {[TmRCo]2+, in CH3CN), 615
{[TmRFe(CH3CN)2(CH3OH)]2+, in CH3OH}. IR (KBr): 2318 (v c n ), 2291
(v c n ). Anal. Calcd for C ^ ^ ^ F g C o i (2.4-2BF4 CH3CN, M = 839.28
g m ol1): C, 51.48; H, 4.08; N, 16.69. Found: C, 51.24; H, 4.04; N, 16.44.
[(tape)Fe(CH3CN)2](BF4)2 (2.5-2BF4)
0.7 ml of 10% 1,2-bis(diallylphosphino)ethane(tape) (0.27 mmol) in THF
was added dropwise to a solution of 2.2*2BF4 (0.21 g, 0.25 mmol) in 20
ml CH2CI2 with stirring in room temperature. The colour of reaction
mixture changed to pale brown. The volatiles were removed in vacuo and
the solid residue dissolved in CH2CI2. The solution was filtered and
concentrated. Red crystals of 2.5*2BF4 were obtained by vapour diffusion
of diethyl ether into the filtrate.
48
Chapter 2 Tripodal Trispvrazolvlmethane Iron and Cobalt Complexes
References
1. S. Trofimenko, Chem. Rev. 1993,93,943.
2. H. R. Bigmore; S. C. Lawrence; P. Mountford; C. S. Tredget, J. Chem. Soc., Dalton Trans., 2005, 635-651.
3. K, Uehara; S, Hikichi; M, Akita, J. Chem. Soc., Dalton, Trans., 2002, 3529-3528.
4. D. L. Reger, C. A. Little; G. J. Long, et al. Inorg. Chim. Acta, 2001, 316,65-70.
5. (a) P. Gutlich, in Mossbauer Spectroscopy Applied to Inorganic Chemistry (Ed.:
G. J. Long), Plenum: New York, 1984; Vol. 1, P287; (b) P. Gutlich, A. Hauser, H.
Spiering, Angew. Chem. Int. Ed Engl. 1994, 33, 2024-2044; (c) O. Kaln, J.
Martinez, Science 1998,279,44-68.
6. D. L. Reger, C. A Little; M D. Smith, Inorg. Chem. 2002,41,4453-4460.
7. (a) D. L. Reger, C. A Little; G. J. Long et al, Eur. J. Inorg. Chem. 2002, (5),
1190-1197;
8. A. Macchioni, G. BeUachioma, G. Cardaci, Organometallics 1997, 16, 2139-
2145.
9. G. M. Sheldrick, SHELXTL97Program for the refinement of Crystal Structures, University o f Goettingen, Germany, 1997.
10. D. L. Reger, T. C. Grattan eta l J. Organomet. Chem. 607,2000,120-128.
11. (a) B.J. Hathaway, A.E. Underhill, J. C. S. 1960, 3705-3711; (b) B.J. Hathaway,
A.E. Underhill, J. C. S. 1962,2444-2449.
12. L. Maier, Helv. Chim. Acta 1966, 49, 842.
13. N. P. Cooper, Doctoral thesis, 1991, Cardiff University
49
Chapter 3
Cyclobutadienylcobalt Template
Complexes
Chapter 3 Cyclobutadienylcobalt Template Complexes
Abstract
The cationic solvated complex (tetramethylcyclobutadienyl)cobalt-
(trisacetonitrile), [Cb*Co(NCCH3)3]+ (Cb* = q4-C4Me4) (3.1), allows the
stepwise introduction of suitable phosphine precursors to the Cb*Co+
fragment by replacement of the labile acetonitrile ligands. These reactions
give rise to the piano-stool complexes, [Cb*Co(dppeXNCCH3)]+ (3.2),
[Cb*Co(dppe)(PH2Ph)]+ (3.3), [<^*Co(dfppb)(NCCH3)]+ (3.4) and
[Cfo^C^dfppbXPIfePh)]* (3.5), where dfppb = 1 ,2-bis(di-2-fluorophenyl-
phosphino)benzene and dppe = l,2-bis(diphenylphosphino)ethane.
Selected examples behave as alternative templates for the synthesis of P3
macrocycle complex, [Cb*Co{ 1,4-bis(2-fluorophenyl)-7-phenyl-
[b,eJh]tribenzo-1,4,7-triphosphacyclononane} ]+ (3.6), promoted by
KOBu1 in almost quantitative yield. The ring-methyl of the
teramethylcyclobutadienyl ligand of the macrocycle complex (3.6) can be
activated with concomitant coupling to the ortho position of a PC6H4
group of the macrocycle by intramolecular dehydrofluorination to give
the hybrid phosphorus-caibon donor complex, {tj4, kP, kP, KP-Me2C4-
[l,4-bis(2-CH2C6H4)-7-C6H5-|b,eJi]tribenzo-l,4,7-triphosphacyclo-
nonane]-l,2}Co+ (3.7), in which the P3 macrocycle bears two 2-
methylphraiyl bridges to the cyclobutadienyl function.
50
Chapter 3 Cvclobutadienvlcobalt Template Complexes
3.1 Introduction
As mentioned previously, formation of P3 macrocycles with smaller than
12-membered ring-sizes has required the use of the q5-CpRFe+ (R = H or
alkyl) template complexes instead of the (CO)3Cr or (CO)sMo template
fragments. The recent availability of the piano-stool complex,
(tetiamethylcyclobutadienyl)cobalt4risacetonitrile,[1] [Cb*Co(NCCH3)3]+
(Cb* = q4-C4Me4), prompted us to investigate the potential application of
this complex fragment as an alternative template for the synthesis of
small ring P3 macrocycles for several reasons. Firstly, these Cb*CoL3+
species are isoelectronic and isostructural with the successful template
precursors CpRFeL3+, (CO)3CrL3, and (CO^MoL. All have d6 metal
centres with a pseudo octahedral configuration with inert capping ligands
(3 x CO for Cr and Mo, CpR for Fe, Cb* for Co) occupying three facial
positions and the other three positions are occupied by labile monodentate
ligands (eg. 3 x CH3CN or q6-aryl for Mo). In practice, this arrangement
should enable the introduction of phosphine precursors to the metal centre
through replacement of the labile ligands allowing subsequent formation
of P3 macrocycles by a range of P-C bond forming methods. And with
low-spin d6 metal centre the reactions can be easily monitored by NMR
spectroscopy. Secondly, Co(HI) has a smaller radius then Fe(II)
(covalent, 1.28 A for Co and 1.26 for Fepl) and the steric bulk imposed
by the Cb* ligand may help force the phosphine precursor into a close
proximity which facilitates the formation of smaller ring P 3 macrocycles
through intramolecular P-C bond formation reactions. Thirdly, the
monocationic Cb*Co+ fragment should enhance solubilities of
intermediates and P3 macrocycle complex products in most polar
solvents. The importance of solubility is highlighted by
[Cp*Co(NCCH3)3]2+ (Cp* = CsMes) which has been previously
51
Chapter 3 Cyclobutadienylcobalt Template Complexes
investigated in our laboratory as a template for the synthesis of
phosphine precursor complexes in most suitable solvents prevented ring
closure to form the target complexes. Fourthly, the predicted instability of
the Cb* moiety (relative to CpR complexes), which can be reduced by
formed on this template.
The preparation of [Cb*Co(NCCH3)3]+ according to literature1191 methods
is presented in Scheme 1 below.
Scheme 1 Preparation of [Cb*Co(NCCH3)3]+ and related complexes
The reaction of 2-butyne with aluminium chloride gives rise to the
cocyclodimerization complex, Ti1-C4Me4AlCl3. This was reacted with
octacarbonyldicobalt affording [Cb*Co(CO)3]+ which was isolated as a
PF6’ salt. The salt was then converted to Cb*Co(CO)2l by reaction with
Me3NO (to abstract CO) and Nal. The arene complex,
[Cb*Co(benzene)]+, was formed by refluxing the iodo complex in
phosphorus macrocycles.[31 In this case, solubility problems of the
lithium metal,141 oxidised[S| or attacked by both nucleophiles16,71 or
electrophiles[81 may enhance the potential to liberate any P3 macrocycles
Me —— —Me +
A1C13
AICI3 + Co2 (CO) 8
CO
AICI3 CgHg
Co
L’ t v < g > >L' = Bu*NC, P(OMe>3 , py
L = CH3CN
52
Chapter 3 Cyclobutadienylcobalt Template Complexes
benzene in the presence of AICI3 as a Lewis acid catalyst. The target
complex, [Cb*Co(NCCH3)3]+, was finally obtained by either photolysis
of the arene complex or thermal ligand exchange in acetonitrile.
3.2 L igand substitu tion reactions o f [C b*C o(N C C H 3)3]+
Although [Cbs*eCo(NCCH3)3]+ (3.1) has a close steric and electronic
analogy to [CpRFe(NCCH3)3]+, differences between the two species are
apparent. The Co(III) centre has a smaller radius and is a harder Lewis
acid than Fe(II). According to the literature/ 11 when [Cb*Co(NCCH3)3]+
was treated with excess P(OMe)3, the substitution of the third CH3CN
was observed to be slow. In addition, when [CbCo(NCCH3)3]+ was
treated with PMe3 at ambient temperature, the third CH3CN ligand was
not replaced. It is of interest to determine whether other sterically
demanding phosphine precursors could be sequentially introduced to the
Co centre by replacing all the three labile acetonitrile ligands from 3.1.
To test this, 3.1 was sequentially reacted with 1,2-
bis(diphenylphosphino)ethane (dppe) and phenylphosphine (Scheme 2).
Scheme 2 Ligand substitution reactions of [Cb*Co(NCCH3)3]+
Co dPPe
l / j " L L
L = CH3CN
3.1
Treatment of 3.1 in acetonitrile with a molar equivalent of dppe causes
the rapid displacement of two CH3CN ligands affording the
©Co H^ Ph
Ph2P ^ 1 >Ph2X L /
8P = 70
3.2
ph2p Ph8P = 70, -19
3.3
©
PPh:
53
Chapter 3 Cyclobutadienylcobalt Template Complexes
bisphosphine-monoacetonitrile complex 3.2, [Cb*Co(dppe)(NCCH3)]+,
which shows a resonance at 8 70.2 ppm in the 31P NMR spectrum. The
chemical shift is consistent with a 5-membered chelate ring.1101 The
replacement of the third CH3CN ligand was carried out by reacting 3.2
with another molar equivalent of phenylphosphine in CH2CI2 to give the
bisphosphine-monophosphine complex 3.3, [Cb*Co(dppe)(PH2Ph)]+. In
the P{ H} NMR spectrum of 3.3, the resonances of coordinated dppe
and coordinated phenylphosphine are observed at 8 70.2 ppm and 8 -18.6
ppm (w ith1 Jhj> = 336 Hz in the 31P NMR spectrum) respectively.
Complexes 3.2 and 3.3 were crystallized as PF6" salts and characterized
by X-ray crystallography. Their structures are shown in Figure 1 and
Figure 2 respectively.
54
Chapter 3 Cyclobutadienylcobalt Template Complexes
Figure 1 ORTEP plot (30% probability level) of 3.2 (H atoms omitted)
C14C15 ci6.C30C31C10C21
C19
Selected bond lengths (A): Col-Nl 1.930(3) Col-Pl 2.2400(9) Col-P2 2.2086(9) Col-Cl 2.047(3) Col-C2 2.013(3) Col-C3 2.010(3) Col-C4 2.057(3) Selected bond angles (°): N l-C ol-Pl 91.98(8) Nl-Col-P2 90.92(8) Pl-Col-P2 87.15(3)
55
Chapter 3 Cvclobutadienvlcohalt Template Complexes
Figure 2 ORTEP plot (30% probability level) of 3.2 (H atoms on phenyl
and methyl groups omitted)
C35 C34
C33
C31 C32
C43
C44
C53 C54
Selected bond lengths (A): Col-Pl 2.2043(7) Col-P2 2.2116(7) Col-P3 2.2196(8) Col-Cl 2.065(2) Col-C2 2.088(2) Col-C3 2.044(2) Col-C4 2.027(2) Selected bond angles (°): Pl-Col-P2 89.88(3) Pl-Col-P3 96.18(3) P2-Col-P3 98.58(3)
56
Chapter 3 Cvclobutadienvlcobah Template Complexes
The structures of 3.2 and 3.3 confirm the sequential replacement of labile
CH3CN ligands from the starting material, [Cb*Co(NCCH3)3]+. Both
complexes 3.2 and 3.3 have three-legged piano-stool geometries about
pseudooctahedral cobalt. The q4-tetramethylcyclobutadiene ligand
coordinates in a facial fashion to the Co(III) centre by occupying three
coordination sites. Bidentate dppe occupies two cis coordination sites as
expected with the last coordination site being occupied by CH3CN in 3.2
or phenylphosphine in 3.3. The distance of Co to the plane defined by the
cyclobutadiene carbons (plane C4) is 1.752(3) A for 3.2 and 1.782(2) A for 33 which are both larger than that in 3.1 (1.670(10) A}. This is
presumably due to the relatively weak rc-acceptor/good a-donor
properties of acetonitrile ligands in 3.1 which allows for a stronger Co-
Cb* interaction compared to that in 3.2 and 3 . 3 . There is also expected
to be a larger repulsion between the phosphine ligands (in 3.2 and 3.3)
and Cb* than between acetonitrile and Cb* (in 3.1). The average Co-P
bond length in 3.2 is 2.224(1) A, while the Co-N bond length of 1.930(3)
A is less than the average value of 1.950(7) A in 3.1. In 3.3, the plane
defined by the P atoms (plane P3) is roughly parallel to the C4 plane with
a deviation of 3.2°. The distance of Co to the centre of the P3 plane is
1.161(1) A; the C0 -PH2PI1 bond length of 2.219(1) A is similar to those
between Co centre and chelate P donors (aver. 2.207(1) A} in 3.3.
57
Chapter 3 Cyclobutadienylcobalt Template Complexes
3.3 Synthesis o f tribenzannu lated 9-m em bered triphospha-
m acrocycles based on tetram ethylcyclobutadienyl tem plates
We have demonstrated above that both bisphosphine and monophosphine
ligands may be sequentially incorporated into the Cb*Co+ fragment by
replacing the labile acetonitriles from 3.1. Thus the Cb*Co+ fragment
appears suitable for pre-organising the phosphine precursors, placing
them in close proximity. If dppe in 3.3 can be replaced by other
bisphosphines bearing an appropriate functionality, e.g. o-F aryl,
intramolecular ring closure may be encouraged to occur.
Our strategy for the synthesis of tribenzannulated triphospha-
cyclononanes is outlined in Scheme 3, and is based upon the
methodology established in our laboratory using the CpRFe+ template.[1,)
Scheme 3 Synthesis of tetramethylcyclobutadienylcobalt(III)-l,4-bis(2-
fluorophenyl)-7-phenyl-[b,e,h]tribenzo-l,4,7-triphosphacyclononane.
H2PPh
L = CH3CN
5P= 643.1 3.4
KOBut
58
Chapter 3 Cyclobutadienylcobalt Template Complexes
The bisphosphine ligand, l,2-bis(di-2-fluorophenylphosphino)benzene
(dfppb) (Scheme 4) was recently prepared in our laboratory and it was
successfully used in the synthesis cyclopentadienyl-Iron(II)-l,4-bis(2-
fluorophenyl)-7 -phenyl-[b,e,h]tribenzo-1,4,7-triphosphacyclo-
nonane.110,111
Scheme 4 Preparation of l,2-bis(dichlorophosphino)benzene
In a modification of the procedure by Kyba, [121 substituting phosgene by
triphosgene, l ,2 -bis(dichlorophosphino)benzene was prepapred by
reaction of 1,2-diphosphinobenzene and triphosgene. The fluoroaryl
derivative dfppb was subsequently prepared by addition of 1 ,2 -
bis(dichlorophosphino)benzene to a solution of Lifo-F-CyTi], formed by
addition of Bu“Li to 2-bromofluorobenzene in diethyl ether at -100 °C.
The dfppb was isolated as an off-white powder in moderate yield.
Treatment of 3.1 in CH3CN with an equimolar amount of dfppb, causes
an immediate colour change from red to orange and the rapid
displacement of two CH3CN ligands, affording the orange bisphosphine-
monoacetonitrile complex [Cfc*Co(dfppb)(NCCH3)]+ (3.4) in quantitative>5 -I 1 _ _
yield. The progress of the reaction was monitored by P{ H} NMR
spectroscopy (Figure 3), which shows a new peak growing upon addition
of dfjppb at 5 64.3 ppm due to coordinated dfppb with a 5-membered
chelate ring; the 19F NMR spectrum shows two resonances at 8 -96.5 ppm
and 8 -97.1 ppm indicating magnetically inequivalent fluorine atoms (no
P-F coupling resolved). The IR spectrum of 3.4 shows the CN stretch at
2269 cm 1.
BiFLi/o-BrFCgH
59
Chapter 3 Cyclobutadienylcobalt Template Complexes
Figure 3 31P{'H} NMR spectrum of 3.4
*i
The CH3CN ligand in 3.4 is labile enough to be exchanged by another
equimolar amount of phenylphosphine which affords the bis(phosphine)-
mono(phosphme) complex [Cft*Co(dfppb)(PH2Ph)]+ (3.5). No obvious
colour change (following addition of phenylphosphine) was observed,
however, a new resonance at 8 -19.6 ppm in the 31P{1H} spectrum1 1 confirms coordination of phenylphosphine ( Jhj> = 339 Hz in the P
NMR spectrum); and the coordinated dfppb is observed at 8 61.7 ppm
(Figure 4). The 19F NMR spectrum shows two resonances at 8 -96.3 ppm
and 8 -97.2 ppm (no P-F coupling resolved). The IR spectrum of
60
Chapter 3 Cyclobutadienylcobalt Template Complexes
compound 3.5 shows the absence of band due to CN group indicating loss
of CH3CN; die PH bond stretches appear at 2304 and 2335 cm'1.
Figure 4 31P NMR spectrum of 3.5
The cyclisation reaction producing the macrocycle complex [Cb*Co{l,4-
bis(2-fluorophenyl)-7-phenyl-[b,e,h]tribenzo-l,4,7-triphosphacyclono-
nane}]+ (3.6) was performed by addition of KOBu to a solution of
complex 3.5 in THF. The mechanism is presumed to involve base
deprotonation of the coordinated phenylphosphine to produce the
phosphide anion, which then acts as a nucleophile substituting the
I
61
Chapter 3 Cyclobutadienylcobalt Template Complexes
fluorine on the adjacent -P(o-F-C6H4)2 , and resulting in ring-closure
(Scheme 5).
Scheme 5: Base promoted substitution of ortho- fluoride
basePRPR
M ©/
R
PRRHP
M ©
1) Coordinated phosphine and electron withdrawing metal activates
arylfluoride to facile substitution.
2) Coordinated phosphide must remain nucleophilic after deprotonation,
i.e. require 18 electron complexes.
3) The metal centre holds the phosphide nucleophile and ortho fluoro-
phenyl group in close proximity.
For the phosphide to remain nucleophilic, the template complex should
be coordinatively saturated (i.e. an 18 electron complex) to avoid the
phosphide acting as a base to other metal centre. The overall process may
be described as an intramolecular dehydrofluorinative P-C coupling
reaction via nucleophilic aromatic substitution promoted by base. The
substitution will also be assisted if the phosphine is ortho to the leaving
group since it will be coordinate to an electron-withdrawing metal
favouring the SnAt reaction.
After addition of 2 mol equivalents of KOBu1 to a solution of 3.5 in THF,
the colour of the solution immediately changed from orange to red, then
back to orange again in less than one minute, the red colour is interpreted
as due to the intermediate phosphide complex. The 31P NMR spectrum of
the reaction solution was recorded, the chemical shifts of the coordinated
62
Chapter 3 Cvclobutadienvlcobalt Template Complexes
primary phosphine and diphosphine disappeared simultaneously with the
growth of two new peaks significantly downshifted to 5 100.9 ppm and 5
94.8 ppm and with relative intensities of 1:2 respectively (AB2 pattern)
(Figure 5). No P-H coupling was observed indicating these resonances to
be due to tertiary phosphines. The air- and moisture-stable orange
compound 3.6 was isolated as its PF6~ salt in >95% yield. The Mass
spectrum of compound 3.6 showed a molecular ion at m/z = 755 amu
which also confirms the formation of macycocycle complex.
Figure 5 31P{1H} NMR spectrum of complex 3.6
!
63
Chapter 1 Cyclobutadienylcobalt Template Complexes
31P{1H} and 19F{1H} NMR spectra of compound 3.6 appear as
complicated AB2XX' (A, B = P, X = F) spin systems. The P-F coupling
and P-P coupling could not be resolved. The ^F^H } spectrum of 3.6 is
temperature dependent suggesting fluxional processes arising from
sterically hindered rotation of the o-fluorophenyl group around the P-Qpso
bonds (Figure 6). The hindered rotation and conformational dynamics in
solution renders the F atoms inequivalent. At room temperature, 19F {1H}
NMR spectra recorded in (C l^ S O show four broad singlets. At elevated
temperature (70 °C), the fluorophenyl rotation becomes fast and only two
averaged lines are observed. At low temperature, 19F{1H} NMR spectra
recorded in CD2CI2 also show four broad singlets. The presence of two
peaks in the high temperature spectrum suggests another magnetically
distinct species (isomer) in solution. This might due to the conformatiom
of the backbone residing in an inverted position with two aryl group
pointing away from the metal. This system needs further study to resolve
this issue.
Figure 6 Variable temperature !9F{!H} NMR spectra of 3.6
1
20 °C, Dg-DMSO 70 °C, Dg-DMSO 20 °C, CD2C12 -50 °C, CD2 CI2
(a) (b) (c) («D
64
Chapter 3 Cvdobutadienvlcobalt Template Complexes
Complexes 3.4, 3.5 and 3.6 were crystallized as PF6" salt and
characterized by X-ray ciystallography. Their structures are shown in
Figures 7 ,8 and 9 respectively.
65
Chapter 3 Cyclobutadienylcobalt Template Complexes
Figure 7 ORTEP plot (30% probability level) of
[Cb*Q)(dlppbXCH3CN)]+ (3.4) (H atoms omitted)
C35
Selected bond lengths (A): Col-Nl 1.934(2) Col-Cl 2.032(3) Col-C2 2.087(3) Col-C3 2.060(3) Col-C4 2.015(3) Col-Pl 2.2241(8) Col-P2 2.2260(8) Selected bond angles (°): N l-C ol-Pl 95.79(7) Nl-Col-P2 94.46(8) PI-Col-P2 85.85(3)
66
Chapter 3 Cyclobutadienylcobalt Template Complexes
Figure 8 ORTEP plot (30% probability level) of
[Cb*Co(dfppb)(PH2Ph)]+ (3.5) (H atoms on phenyl and methyl groups
omitted)
C14 C13
C12
C11
C30C29
Selected bond lengths (A): Col-Cl 2.023(3) Col-C2 2.078(3) Col-C3
2.104(3) Col-C4 2.053(3) Col-Pl 2.2190(8) Col-P2 2.2139(8) Col-P3
2.2125(7) Selected bond angles (°): Pl-Col-P2 97.79(3) Pl-Col-P3
88.86(3) P2-Col-P3 86.90(3)
67
Chapter 3 Cvclobutadienvlcobalt Template Complexes
Figure 9 ORTEP plot (30% probability level) of 3.6 (H atoms omitted,
FI A and FIB are disordered in a ratio of 0.6:0.4.)
C38(a)
C43 C42C4lC30
45CAOJP2S^
(b)Selected bond lengths (A): Col-Cl 2.015(5) Col-C2 2.016(4) Col-C3 2.061(4) Col-C4 2.054(4) Col-Pl 2.1534(12) Col-P2 2.1615(12) Col- P3 2.1643(12) Selected bond angles (°): Pl-Col-P2 88.10(5) Pl-Col-P3 87.74(4) P2-Col-P3 88.35(4) C(9>P(l)-Co(l) 123.13(17) C(15>P(1)- Co(l) 109.72(15) C(45>P(l)-Co(l) 109.70(14) C(22)-P(2)-Co(l)126.71(14) C(28)-P(2)-Co(l) 109.23(14) C(20)-P(2>Co(l) 108.93(15)
68
Chapter 3 Cvclobutadienvlcobah Template Complexes
C(34>P(3>Co(l) 126.97(14) C(33)-P(3)-Co(l) 109.11(13) C(40)-P(3> Co(l) 109.16(13)The structures of 3.4 and 3.5 are quite similar to those of 3.2 and 3.3
respectively. Again both complexes 3.4 and 3.5 have three-legged piano-
stool geometries about a pseudooctahedral cobalt centre. The q4-
tetramethylcyclobutadiene ligand coordinates in a facial fashion to the
Co(III) centre by occupying three coordination sites. The distances of Co
to the plane defined by the cyclobutadiene ring carbons (plane C 4 ) are
1.773(3) A in 3.4 and 1.789(3) A in 3.5, both of which are larger than that
(1.670(10) A} in l .[61 As mentioned before, this is presumably due to the
weak x-acceptor/good o-donor property of CH3CN ligand in 3.1 allowing
for a stronger Co-Cb* interaction, 161 as well as the larger repulsion
between the phosphine ligands and Cb*. The Co-N bond length of
1.934(2) A in 3.4 is comparable to that in 3.2 (1.930(3) A}, both are less
than the average value of 1.950(7) A in 3.1. The average Co-P bond
length in 3.4 is 2.225(8) A; the Co-PH2Ph bond length (2.190(1) A} in
3.5 is similar to the values between Co and chelate P donors (aver.
2.213(1) A}. In 3.5, the dihedral angle between C4 and P 3 planes is
8.8(1)°, the distance of the Co to P 3 plane centroid is 1.247(1) A.
Structure 3.6 confirms the formation of the triphosphamacrocycle which
sandwiches the Co centre together with the q4-cyclobutadiene. The
macrocycle ligates to the metral centre via three fused 5-membered rings
as a tridentate crown. The roughly parallel P 3 and C 4 planes have a
dihedral angle of 4.1(1)°. The distance of Co to the C4 plane centroid is
1.762(5) A, and to the P 3 plane centroid is 1.288(1) A. The Co-P bond
lengths in 3.6 are 2.153(1) A, 2.161(1) A, 2.164(1) A which are
significantly shorter than those in 3.5 (aver. 2.215(8) A}, the complex
prior to ring closure. This exemplifies the macrocycle coordination effect
leading $0 strpng bonding between the metal centre and the macrocyclic
69
Chapter 3 Cvclobutadienvlcobalt Template Complexes
ligand. This strong bonding is also manifest in the 31P NMR spectrum
causing the resonances to be shifted downfield in comparison to 3.5. The
fluorine positions in the structure are disordered and thus it is not possible
(from structural data) to identify definitively the two phosphorus atoms
bearing the o-fluorophenyl substituents. However, two Co-P bond lengths
are longer (2.161(1) A, 2.164(1) A} than the third (2.153(1) A}. It is
likely that these two longer contacts are to the o-fluorophenyl bearing
phosphines. This argument is supported by the P( H} NMR spectrum
since an AB2 pattern is observed where 6 A is more deshielded, consistent
with a slightly stronger Co-P interaction, and 8 B is more shielded with a
slightly weaker Co-P interaction due to the electron withdrawing
influence of the o-fluoro group.
In 3.4, the o-phenylene carbons and two chelating phosphorus atoms are
roughly co-planar. This plane and the trigonal plane defined by the two
chelating P and Co atoms has a dihedral angle of 10.6(1)°. In 3.5, the
dihedral angle of the corresponding two planes changes to 27.2(1)° This
could be due to a larger replusion between the phenylphosphine and the
dfppb ligand compared to that between acetonitrile and phenylphosphine
in 3.4. The average corresponding dihedral angle between these planes in
3.6 is 9.5(1)°. In 3.4, the dfppb coordination bite angle is 85.83(3)°, The
P—P non-bonded distance is 3.031(1) A. In 3.5, this dfppb bite angle
angle is 86.90(3)°, the other P-Co-P angles are 88.86(3)° and 97.79(3)°,
and the P—P non-bonded distance associated with the 5-membered
chelate ring is 3.044(1) A; and the others are significantly longer at
3.102(1) A and 3.340(1) A. In 3.6, the P-Co-P angles of the three fused 5-
membered chelate rings are 88.10(5)°, 87.74(4)°, 88.35(4)°, the average
P—P distances is 3.000(1) A.
70
Chapter 3 Cvclobutadienvlcobah Template Complexes
The methyl groups of cyclobutadiene in 3.6 are displaced out of the C4
ring plane away from the triphosphamacrocycle as is also observed in
complexes 3.2, 3.3, 3.4 and 3.5 this might be presumably due to
combination of steric repulsion and electronic factor.1131 The
intramolecular repulsion between the Cb* ligand and the terminal phenyl
or fluorophenyl group in 3.6 causes the terminal phenyl or fluorophenyl
to be bent away from the cobalt atom with the C o-P -C exo (exocyclic
phenyl or fluorophenyl group) angles expanded to 123.1(1)°, 126.7(1)°,
126.9(1)° respectively, whereas the Co-P-Cnng angles are close to
tetrahedral (aver. 109.5(1)°}. It is likely that the two larger C o-P-C exo
angles belong to the more sterically demanding fluorophenyl groups,
although due to crystallographic disorder, this could not be confirmed
structurally.
71
Chapter 3 Cvclobutadienvlcobah Template Complexes
3.4 Intramolecular C-C bond coupling between phosphine
macrocycle and tetramethylcyclobutadiene
Since the fluorine atom is generally a good leaving group for SnAt
reactions and the dehydrofluorination is stoichiometric in base, addition
of 2 mol equivalents of strong base to a solution of 3.5 in THF results in
removal o f two fluorine atoms and 3.6 is formed in almost quantitative
yield. With excess base however, two cis Cb* ring methyl C-H bonds and
the two remaining o-aryl C-F bonds were activated which leads to an
intramolecular C-C coupling between the cyclobutadienyl ligand and the
P3 macrocycle ligand. The selectivity of the reaction in which 3.6 is
formed preferentially to intramolecular C-C coupling indicates the PH
protons in 5 are more acidic than the CH protons in the Cb* ligand.
Addition of more than 2 equivalents of KOBu1 to a solution of 3.5 in THF
gave rise to the air- and moisture- stable complex 3.7 (vide supra). Complex 3.7, {Ti4-Me2C4-[l,4 -bis(2 -CH2C6H4)-7 -C6H5-[b,e,h]tribenzo-
1,4,7-triphosphacyclononane] -1,2} Co+, can also be rapidly and almost
quantitatively formed by adding KOBu to the solution of 3.6 in THF. It is
likely that 3.7 was formed via C-H bond activation and concomitant
intramolecular C-C bond formation by nucleophilic substitution of
fluoride and constitutes an unprecedented coupling of a
tetramethylcyclobutadienyl ligand with a fluoroaryl phosphine (Scheme
6).
72
Chapter 3 Cyclobutadienvlcobalt Template Complexes
Scheme 6: Cb* ring methyl activation and concomitant coupling reaction
©
F Co Co c h 2--H / / S T I \ F'
p" i p—(Q >
KOBo ^ l-/ r
Co
8 P = 101,94
3.6-F‘
Co Co-HF
KOBu
3.7 3.7a
The mass spectrum of 3.7 showed peaks at 712 (M-2H) and 711 (M-3H)
amu. The 31P{1H} NMR spectrum of complex 3.7 (Figure 10) recorded
in CD2CI2 shows two resonances shifted downfield relative to 3.6 at 5
106.3 ppm and 5 102.2 ppm with relative intensities 2:1 (A2B pattern). P-
P coupling could not be resolved. The NMR spectrum of 3.7
showed solely a doublet at 6-71.6 ppm and 6 -74.1 ppm (]JP}f = 710 Hz)
due to the PF6 counter anion. The two methylene protons of the CH2
bridges are diastereotopic and inequivalent. One appears as a doublet (6
3.08 ppm) in the lU NMR spectrum through coupling to its geminal
neighbour (2Jh3 =16.8 Hz) while the other is a double of triplets (5 3.57
ppm) through further coupling to one aromatic proton and a phosphorus.
In the 13C{ H} NMR spectrum, the carbons in the Cb* ring are
chemically inequivalent and show two peaks at 6 86.6 ppm and 6 78.3
ppm, a resonance at 6 28.3 ppm is assigned to the methylene carbon and
the methyl carbon gives rise to a resonance at 5 9.3 ppm.
73
Chapter 3 Cvclobutadienvlcobalt Template Complexes
Figure 10 31P{]H} (a) and NMR (b) spectra of complex 3.7
aa
UM 1000 96l0
(a)
3.6 3.4 3.2 3.0
(b)
„ 5 *1 -The ring methyl C-H bond activation of an
pentamethylcyclopentadienyl ligand and an t|6-arene ligand, followed by
intramolecular coupling to the o-aryl position of a fluorophenyl
phosphine ligand in transition metal complexes has been reported.114"181
74
Chapter 3 Cvclobutadienvlcobalt Template Complexes
As far as we know, this is the first example of the ring methyl C-H bond
of an ri4-tetramethylcyclobutadienyl ligand, undergoing a similar
intramolecular coupling to a phosphine ligand bearing an o-fluorophenyl
group. The activation of the ring methyl of Cb* provides some
opportunities further to investigate the reactions of the activated
nucleophile intermediate with other electrophiles in the future.
The precise nature of the ring methyl activation of t j 5 - CsMes or t)4-
C4Me4 or ij6-arene ligands, followed by intramolecular or intermolecular
coupling reactions remains speculative. Firstly, there is an unexpected
sensitivity of the reactions to the nature of the organometallic base used.
In some cases, only KOBu1 was successful, LiNPr2 or other bases are less
effective*15’191 and in some of Saunders’ cases,*141 no base is necessary.
Secondly, in the case of i]5-Cp*, single or multiple alkylations occur
depending upon the metal system.*191 The reaction of 3.6 to give 3.7 can
be simply viewed as an overall addition-elimination reaction or a
nucleophilic aromatic substitution. We suggest a mechanism similar to
that proposed by Saunders*14,16-191 and Hughes*151 where deprotonation of a
methyl group on the Cb* ring occurs to form a neutral intermediate
bearing a methylene carbanion (Scheme 6 ). The resulting nucleophilic
methylene caibon attacks at the o-position of the fluorophenyl group
resulting in fluorine elimination and the formation of the new C-C bond.
The cobalt centre holds the cis ring methyl of the Cb* ligand and the o- fluorophenyl group of the P3 macrocycle ligand in close proximity,
facilitating the nucleophilic attack. The dehydrofluorinative C-C coupling
reaction is fast on the NMR timescale since we could not detect (NMR)
or isolate intermediate 3.7a, even when a deficiency of base was added to
the solution of 3.6.
75
Chapter 3 Cvclobutadienvlcobalt Template Complexes
Complex 3.7 was crystallized as the SbF6~ salt by anion exchange and
characterized by X-ray crystallography (Figure 11). The cobalt in 3.7 also
has a pseudooctahedral geometry similar to complexes 3.1 to 3.6. There are
however obvious structural differences between 3.7 and the other complexes
reported herein. The cobalt centre is encompassed by the two parts of a
hybrid triphosphamacrocycle-cyclobutadienyl ligand. The cyclobutadiene
part occupies three fac coordination sites, the P3 macrocycle part occupies
the other three mutually cis sites. Although the two methylene carbon atoms
and a ring carbon atom of the cyclobutadiene in 3.7 are disordered, the
structure still clearly shows the triphosphamacrocycle linkage to the cyclobutadiene group via two cis adjacent positions of the cyclobutadiene
ring. The hybrid macrocyclic phosphine-cyclobutadiene ligand binds the Co
centre strongly acting as a 12 electron formally hexadentate donor. The
distance of 1.716(8) A from Co to the cyclobutadiene ring centroid is shorter
than that in 3.6 (1.762(5) A}. All three Co-P distances in 3.7 become much
shorter compared to those in 3.6 after removal of the electron withdrawing
fluorine atoms. The two P atoms bearing the 0 C6H4CH2 linkage to the Cb*
fragment have Co-P bond lengths of 2 .1 2 0 (1) A and 2.125(1) A. The
remaining uncoupled P atom has a Co-P bond length of 2.122(1) A. In 3.7,
the average P-Co-P bite angles are approximately 89.77(7)°, the average
P—P distance of 2.996(1) A is almost identical to that in 3.6 (3.000(1) A} which reflects the rigidity of the tribenzannulated backbone. The Co-P-C(Cl
and C20 on the coupled methylenylphenyl) angles (aver. 119.2(1)°} are
smaller than that of the Co-P-C33 (124.4(1)°}. The Co-P-Cl and Co-P-C20
angles are both significantly smaller than those in the uncoupled precusor
(3.6) due to the constraints imposed as a result of coupling. All the Co-P-
Cexocyciic phenyl angles in 3.7 are larger than those of the backbone Co-P-Cnng
angles (aver. 108.3(1)°} due to intramolecular repulsion as observed in
complex 3.6.
76
Chapter 3 Cvclobutadienvlcobalt Template Complexes
Figure 11 ORTEP plot (30% probability level) of 3.7 (H atoms and
disordered C7B, C11B, C12B, C14B omitted)
&C12A• C14A C16
€35
(b)Selected bond lengths (A): C ol-Pl 2.1200(17) Col-P2 2.1220(19) Col- P3 2.1259(18) C0 I-C8 1.985(6) Col-C9 2.000(8) C ol-C lla 1.989(10) Col-C13 1.994(7) Selected bond angles (°): Pl-Col-P2 90.10(7) P l-C ol- P3 89.37(6) P2-Col-P3 89.84(7) C(l>P(l)-Co(l) 120.0(2) C(26)-P(l)- Co(l) 107.70(19) C(44>P(l)-Co(l) 109.47(19) C(33)-P(2)-Co(l)124.4(2) C(39)-P(2)-Co(l) 108.3(2) C(32)-P(2)-Co(l) 107.8(2) 0(21 > P(3)-Co(l) 107.9(2) C(20)-P(3)-Co(l) 119.2(2) C(27>P(3>Co(l)108.8(2)
77
Chapter 3 Cvclobutadienvlcobalt Template Complexes
3.5 Cyclic Voltammetry
Cyclic voltammetry (CV) gives information about electrochemical
process which may be relevant to understanding chemical reactions. By
comparison of die redox potentials of the electroactive species obtained
from CV, we can get the insight into the influences that various ligands
may have upon the metal redox centre and reduction/oxidation resistance
of redox active centres. This information could help in the selection of the
most suitable reagents to oxidise or to reduce the redox active species.
Since we intend to study the liberation of the P3 macrocycle from
complexes 3.6 or 3.7, by reducing or oxidising the kinetically inert
species in order to disrupt the d6/18 electron configuration of the Co
centre and so to form relatively kinetically active intermediates with a5 7d /17 or d /19 electron configuration. CV measurements were performed
in this regard (Table 1, Figure 12 and Figure 13).
Table 1 Cyclic voltammetric data for the oxidation of complex 3.6 and
3.7 in CHsCN*
Complex process V00 Ep^CV) AEp(mv) j ox/- red Ip /lp
3.6 1 0.407 Irreversible
2 0.610 0.535 75 0.93
3.7 1 0.614 0.529 85 0.99
a) Potential is referenced to Cp2Fe+/Cp2Fe; scan rate: 400mv/s; Ep0X =
oxidation peak potential; Ep1®*1 = reduction peak potential; AEp =
separation in reduction and oxidation peak potentials; Ipox = oxidation
peak current; Ip”* = reduction peak current
78
Chapter 3 Cvclobutadienvlcobalt Template Complexes
Figure 12 oxidative cyclic voltammogram for complex 3.6
.20x101.10x10'
1.10x10'0.250 0.750
E/ V
Figure 13 oxidative cyclic voltammogram for complex 3.7
).03x10** }.05x10* ).08x10'*
0.250 0.500 0.750 1.000 1.250 1.500E / V
In acetonitrile, both 3.6 and 3.7 show no reduction within the solvent
limit (-2.5 V). Thus the P3 macrocycle ligand/Cb*Co+ fragment forms a
very stable complex against reduction. Complex 3.6 shows two oxidation
processes, one irreversible taking place at 0.407 V, another reversible at
0.610 V (vs Cp2Fe+/Cp2Fe couple). Complex 3.7 shows one reversible
oxidation at 0.614 V (vs Cp2Fe+/Cp2Fe couple). The first irreversible
oxidation process for 3.6 is explained as a ligand based oxidation and
coupled to the chemically irreversible formation of complex 3.7 by
electrochemical oxidation. Hence the second oxidation process of 3.6
occurs at the same voltage and has the same shape as complex 3.7 which
is assumed to be a metal based single electron oxidation from the 18
electron Co(III) species to a 17 electron Co(IV) species. From the CV, it
seems that the reaction of 3.6 with KOBu affording 3.7 involves a
complicated electron transfer process. Further studies are required in
order to distinguish one or two electron process in order to gain
mechanistic information of the transformation of 3.6 to 3.7.
79
Chapter 3 Cvclobutadienvlcobalt Template Complexes
3.6 Ligand liberation studies
Several routes have been tried in an attempt to liberate the 1,4,7-
tribenzannulated triphosphacyclononane ligand from complex 3.6,
although without success. The general description of these studies is
presented here.
The basic principle for liberation is initially to transform the kinetically
inert 18 electron complex with a d6 cobalt centre to a more kinetically
active 17 or 19 electron complex by oxidation or reduction. The addition
of 2 molar equivalents of LiHB(C2H5) 3 to a solution of 3.6PF6 in THF
caused the solution to immediately change colour from yellow to red. An
intermediate which is soluble in hydrocarbon solvents was isolated and
shows a upfield chemical shift in the 3,P{1H} NMR spectrum compared
to 3.6, no PF6' was observed. The gradual degeneration of the
intermediate to show a spectrum similar to that of 3.7 however caused us
to abandon this course of investigation.
Reductive cleavage of ferrocene to metallic iron and cyclopentadiene by
lithium has been reported.1203 Sekiguchi et al also reported the synthesis
of the cyclobutadiene dianion from CpCocyclobutadiene complexes by
reaction with metallic lithium.1213 Thus, reaction of 3.6 PF6 in THF
solution with metallic lithium was also tried. After stirring overnight, the-j ■% i __ ___
solution showed no resonance in the P{ H} NMR spectrum, although,
we have not been able to identify any products from this reaction.
Oxidation of 3.6*PF6 was carried out in acetonitrile solution with eerie
ammonium nitrate. The reaction proceeds quite slowly compared to the
Fe analogue. After stirring overnight, the 31P NMR spectrum shows no
resonances, presumably due to the formation of a paramagnetic complex.
80
Chapter 3 Cvclobutadienvlcobalt Template Complexes
We have been unable to isolate any identified products during following
work.
The attempt to substitute the otho fluorine atom in the PQfLjF groups by
KCN was investigated. Upon heating a solution of 3.6PF6 with excess
KCN in phenyl nitrile (160°, >12 h), no reaction was observed.
Gleiter reported cycloaddition reactions of CpCo-stabilized
cyclobutadiene derivatives with triple bonded species X=Y, to yield
arenes or pyridines under vigorous conditions.*221 No similar reaction was
observed by treating complex 3.6 with phenyl nitrile at 200 °C for more
than 1 2 hours however.
81
Chapter 3 Cvclobutadienvlcobalt Template Complexes
3.7 Conclusions
The lability of acetonitrile ligands of the cationic [Cb*Co(NCCH3)3]+
complex makes the electrophilic Cb*Co+ fragment accessible to
bisphosphine and monophosphine ligands. By replacing the acetonitrile
ligands from [Cb*Co(NCCH3)3]+ sequentially with bisphosphine and
monophosphine ligands, the precursor complex, [Cb*Co(dfppb)(PH2Ph)]+
(3.5) is available which undergoes ring closure to produce a 9-membered
1,4,7-triphosphorus macrocycle complex (3.6) with a tribenzannulated
backbone via intramolecular dehydrofluorinative coupling reactions using
KOBu1 as a HF-trapping reagent. This makes the Cb*Co+ unit the first
alternative non-Fe template for the synthesis of P3 macrocycles with
small ring systems.
Surprisingly, in the presence of KOBu1, the Cb* ring methyl of complex
3.6 undergoes C-H bond activation and concomitant coupling to the ortho position of a coordinated PC6H4 group. This produces complex 3.7
bearing a 1 2 electron donor ligand with hybrid phosphine and
cyclobutadienyl groups. The significantly shorter Co-P bond lengths in
3.6 and 3.7 suggest a very good ligating behaviour of the P 3 macrocycle
to the metal centre that renders corresponding complexes of 3.6 and 3.7
very stable and resistant to reduction (CV).
82
Chapter 3 Cvclobutadienvlcobalt Template Complexes
3.8 Experimental Section
General experimental information: All reactions were performed under
nitrogen atmosphere with standard Schlenk glassware, vacuum
techniques or Glovebox unless otherwise noted. The solvents were dried
and degassed by refluxing over standard drying agents and distilled
immediately prior to use. Infrared spectra were recorded on a Perkin
Elmer 1600 spectrophotermeter using KBr pellets for solid samples. Mass
spectra were carried out on a VG Platform II Fisons mass spectrometer.
UV photolyses were carried out using a Hanovia 125 W mercury
discharge lamp (254 nm). The NMR spectra were recorded on a Bruker
DPX-400 instrument at 400 MHz (]H) and 100 MHz (13C) or a Jeol
Lamda Eclipse 300 at 121.65 MHz (31P), 300.52 MHz (*H), 75.57 MHz
(13C), 96.42 MHz (nB) and 282.78 MHz (19F). All chemical shifts are
quoted in units of 5 ppm. *H and 13C NMR chemical shifts are relative to
solvent resonance, 31P chemical shifts are relative to 85% external H3PO4
( 8 = 0 ppm), nB chemical shifts are relative to external BF3.OEt2 ( 8 = 0
ppm), 19F chemical shifts are relative to external CFCI3 ( 8 = 0 ppm).
Elemental analyses were performed by the Warwick Analytical Service.
X-ray diffraction data collection was carried out on a Bruker Kappa CCD
diffractometer at 150(2) K with Mo Ka irradiation (graphite
monochromator). Empirical absorption corrections were performed using
equivalent reflections. For the solution and refinement of the structures,
the program package SHELXL 97 was employed.1231 H atoms were
placed into calculated positions and included in the last cycles of
refinement. Complexes 3.2 to 3.6 were aystallised as its PF6 salts,
complex 3.7 as SbF6 salt. The two F atoms in complex 3.6 are refined and
distributed between the three phenyls with a disordered model of
1:0.6:0.4. In complex 3.7, the positions of two CH2 carbons, a CH3 carbon
83
Chapter 3 Cvclobutadienvlcobalt Template Complexes
and a C atom from the Cb* ring are refined isotropicaUy in a disordered
mode of 0.6:0.4. Crystal structures and refinement data are collected in
Appendix B and supplementary CD.
Electrochemical measurements: Cyclic voltammetry was carried out on
a Windsor PG system versus an Ag/Ag+ reference electrode, a glassy-
carbon working electrode and a platinum wire as auxiliary electrode. The
measurements were recorded on ca 1.0 mM solutions of the compounds
in acetonitrile (0.1 M BU4NPF6 as supporting electrolyte). 1.0 mM
ferrocene solution in the same solvent was used as reference redox couple
and all potentials are quoted relative to Fc+/Fc. Solutions were purged
with nitrogen before the measurements. All other chemicals were
obtained from commercial sources and where appropriate, dried over
molecular sieves and degassed by repeated freeze-thaw cycles.
M aterials: [(CM ^ofNCCUs^PFe11 (3.1PF6) and phenyl
phosphine1241 were prepared according to literature methods.
1 . 2 -bis(dichlorophosphmo)benzenell0]This compound was prepared by a modification of the procedure of
Kyba, [121 substituting phosgene by triphosgene. To a solution of 24.4 g
(63 mmol) triphosgene in 250 ml dichloromethane was added a solution
of 6.7 g (47 mmol) 1,2-diphosphinobenzene at -78 °C. After addition was
complete the reaction mixture was left to warm up to room temperature,
the solvent was evaporated and the residue distilled in high vacuum (at
around 0.1 mmHg, 135 °C) to give a clear liquid, 4.9 g (37%).
1.2 -bis(di-2 -fluorophenylphosphino)benzene (dfppb)[10]
89 ml of a 2.13 M solution of LiBu11 (0.19 mol) in diethyl ether was
diluted with 200ml ether and cooled to -100 °C. To this was added
dropwise a solution of 20.6ml (33 g, 0.19 mol) of 2-bromofluorobenzene
in 30 ml ether. After the addition was complete the solution was stirred
84
Chapter 3 Cvclobutadienvlcobalt Template Complexes
for 30 minutes and then a solution of l,2-bis-(dichlorophosphino)benzene
(13.2 g, 0.047 mol) in 25 ml ether was added. The mixture was left to stir
for 1 h at low temperature and then left to warm up to room temperatue
overnight. An aqueous solution of NH4CI was added, the organic phase
separated and concentrated and the residue crystallized from ethanol.
Yield: 33%.
[(T|4-C4 Me«)Co(dppeKNCCH3)]PF6 (3.2PF6)
To a solution of [(C4Me4)Q>(NCCH3)3]PF6 (3.1PF6)(0.10 g, 0.23 mmol)
in 20 ml CH3CN, dppe (0.09 g, 0.23 mmol) was added with stirring. The
colour of the deep red solution changed to yellow-brown immediately.
The solution showed a resonance at 8 70.2 ppm in the 31P{1H} NMR
spectrum. The reaction mixture was filtered then concentrated. Orange
crystals of 3.2 *PF6 were obtained by diffusion of diethyl ether vapour into
the solution. Yield: 98%.
[(Tl4-C4Me4)Co(dppe)(PH2Pli)]PF<(3.3PF6)
To a solution of (3.2-PF6) (0 .1 0 g, 0 .1 2 mmol) in 2 0 ml CH2CI2,
phenylphosphine (17mg, 0.15mmol) was added. After stirring for 20
minutes, the 31P NMR spectrum showed two resonances at 5 70.2 ppm
and 8 -18.6 ppm. The volatiles were then removed in vacuo. The solid
residue was triturated with petroleum ether then dissolved in CH2CI2,
filtered and concentrated. Orange crystals of 3.3 PF6 were obtained by
diffusion of diethyl ether vapour into the filtrate. 31P{1H} (CDCI3): 70.2, -
18.6 ( ^ = 336 Hz). Yield: 98%.
[(Tl4-C4Me4>Co(o-C6H4 F)2PC(>H4P(u-C6H4 F)2(NCCH3)]PF6 (3.4 PF6)
To a solution of 3.1PF6 (0.10 g, 0.23 mmol) in 20ml C H 3 C N , dfppb
(0.12 g, 0.23 mmol) was added with stirring. The colour of the red
solution changed to orange immediately. 31P{1H} NMR spectrum showed
85
Chapter 3 Cvclobutadienvlcobalt Template Complexes
a growth of a resonance at 5 64.3 ppm. The reaction mixture was filtered
then concentrated. Orange crystals of 3.4-PF6 were obtained by diffusion
of diethyl ether vapour into the solution. Yield: 98%. 31P{1H} (CD2CI2):
64.3 (dfppb), -143.9 (sep PFfc 'j P>F = 711 Hz). i9F{'H} (CD2CI2): -96.5
(dfppb), -97.1 (dfppb), -73.4(d, PF«). 13C{’H} (CD2CI2): 85.1 (C«Me«),
8.2 (Q M e^, 1.4 (MeCN). 'H (CD2CI2): 7.6 to 6.4 (20H, m, H ^,), 0.91
( 1 2 H, s, GtMe4), 0.79 (3H, s, CHjCN). IR: 2269 cm'1 (CN) Anal. For
C40H35N1F10P3C01 (M = 871.56 g mol'1): Calcd (%), C, 55.12; H, 4.05; N,
1.61. Found (%), C, 54.78; H, 4.05; N, 1.51.
[(Ti4-C4Me4)Co(o-QH4F)2PC6H4P(t>-C<H4F)2(PH2Ph)|PF6 (3.5PF6)
To a solution of 3.4*PF6 (0.20 g, 0.23 mmol) in 20ml CH2CI2,
phenylphosphine (0.03 g 0.27 mmol) in toluene was added. After stirringHI _ __
for 15 minutes, the P NMR spectrum showed two resonances at 8 61.7
ppm and 5 -19.6 ppm. The reaction mixture was filtered and the solvent
removed in vacuo. The residue was triturated with diethyl ether to remove
excess amount of phenylphosphine and the yellow powder dissolved in
CH2CI2 and concentrated. Orange crystals of 3.5*PF6 were obtained by
diffusion of diethyl ether vapour into the concentrated solution during
several days. Yield: 98%. 31P (CD2CI2): 61.7, -19.7 (t, 'j P,H = 336 Hz), -
143.8 (sep, PF& 'jP>F = 711 Hz). I9F{*H} (CD2CI2): -96.3 (dfppb), -
97.2(dfppb), -73.3 (d, PF6). ,3C{'H} (CD2CI2): 83.7 (GiMe*), 6.72
(G,Me4). *H (CD2CI2): 7.4 to 6.7 (m, 25H, H ^ ), 4.0 (dm, 2H, PH, 'j PJ1 =
336 Hz), 0.86 (s, 12H, C ^e-,). IR: 2304 cm'1 (PH), 2335 cm 1 (PH).
Anal, for C ^ F ^ C o , (M = 940.61 gmol'1): Calcd (%), C, 56.19; H,
4.18. Found (%), C, 55.38; H, 4.04.
86
Chapter 3 Cvclobutadienvlcobalt Template Complexes
[q4-C4M ei]Cobalt(III)-l,4 -bis(2 -fluorophenyl)-7 -phenyl-[b,e,h]tri-
benzo-1,4,7-triphosphacyclononane hexafluorophosphate (3.6*PF6)
To a solution of 3.5PF6 (0.27 g, 0.29 mmol) in 20ml THF was added
KOBu1 (65 mg, 0.58mmol). The solids were dissolved quickly and the
colour of the solution changed immediately from orange to red then back
to orange again. After stirring for 15 minutes, the 31P NMR spectrum
showed two resonances at 5 102.0 and 8 93.9 ppm with intensities 1:2.
The solvent was then removed in vacuo. The residue was dissolved in
dichloromethane and filtered. The solvent was removed in vacuo and
replaced by acetonitrile, orange crystals of 3.6*PF6 'CH3CN were obtained
by diffusion of diethyl ether vapour into the solution. Yield: 95%.
3,P{*H} (CD2CI2): 100.9 (m), 94.8 (m), -143.0 (sep, PF6, = 711 Hz).
-96.5, -96.8, -97.2, -97.4 (CD2CI2 2 , -50 °C); -95.9, -96.3, -96.6
(CD2CI2, 2 0 °C); -96.0, -96.1, -96.6, -96.7 (DMSO-D6, 2 0 °C); -96.4, -
96.9 (DMSO-D6, 70 °C). ,3C{*H} (CDjClj): 82.4 (QMe4), 6.2 (QMe*).
'H (CD2CI2): 7.6 to 7.1 (m, 25H, H ^ ), 1.8 (s, 3H, CHjCN), 0.7 (s, 12H,
G|Me«). M.S.(APCI): 755(M+). Anal, for C46H40N1F8P4Co1 (M = 941.65
gmof1): Calcd (%), C, 58.67; H, 4.28; N, 1.49. Found (%), C, 58.59; H,
4.19; N, 1.32.
{ij4, kP, kP, KP-Me2C4-(l,4-bis(2-CH2C6H4)-7-C«H5-[b,e,h]tri-benzo-
l,4,7-triphosphacyclononane]-l,2}Co(III) hexafluorophosphate
(3.7-PF6)
To a solution of 3.6*PF6 (0.10 g, 0.1 mmol) in 20ml THF was added
KOBu1 (24 mg, 0.2 mmol). The solids dissolved quickly and the colour of
the solution changed immediately from orange to red, then back to 1orange. The P{ H} NMR spectrum showed two new resonances at 8
106 and 8 102 ppm with intensities 2:1. The solvent was removed in
vacuo and the residue was dissolved in dichloromethane and filtered. The
87
Chapter 3 Cvclobutadienvlcobalt Template Complexes
solvent was removed in vacuo and the residue dissolved in acetonitrile.
Orange crystals of 3.7*PF6 were obtained by vapour diffusion of diethyl
ether into the solution. Yield: 95%. 31P{1H}(CD2Cl2): 106.3 (m), 102.2
(m), -143.1 (sep, PF6, = 712 Hz). I9F{'H} (CD2C12): -72.9(d,
PF6).13C{'H} (CD2CI2): 84.6 (Cb ring), 78.3 (Cb ring), 26.5 (methene),
7.0 (ring methyl). *H (CDjCfe): 7.3 to 8.4 (m, 25H, H ^ ), 3.6 (m, 2H,
CH2), 3.08 (d, 2H, CH2, 2Jhh = 16.8 Hz), 1.08 (s, 6 H, Me2C4)
M.S.(APCI): 712 (M-2H), 711 (M-3H).
88
Chapter 3 Cvclobutadienvlcobalt Template Complexes
References
1. G. E. Herberich, A. R Kudinov et al, Eur. J. Inorg. Chem. 2002,2656-2663
2. J. G. Speight, Lange's Handbook of Chemistry 16th Ed. McGraw-Hill, New York,
2005.
3. R. Haigh, Doctoral Thesis, 2002, Cardiff University
4. A. Sekiguchi; T. Matsuo; H. Watanbe; J. Am. Chem. Soc. 2000,122, 5652-5653.
5. C. Schaefer, R Gleiter, F. Rominger, Organometallics 2004, 23, 2225-2227.
6. (a) A. Efraty, Chem. Rev. 1977,77,691. (b) P. M. Maitlis, A. Efraty, M. L.
Games, J. Am. Chem. Soc. 1965, 87, 719
7. M. Crocker, M Green, A. G. Orpen, H.-P. Neumann, C. J. Schaverien, J. Chem. Soc., Chem. Common. 1984,1351. (b) M. Crocker, M Green, K. R Nagle, A. G.
Orpen, H.-P. Neumann, C. J. Schaverien, Organometallics 1990, 9,1422.
8. M. V. Butovskii, U. Englert, G. Herberich, K. Kirchner, U. Koelle,
Organometallics 2003,22,1989-1991.
9. P. L. Pauson et al, J. Chem. Soc. Dalton Trans. 1987, 2757-2760
10. P. Garrou, Chem. Rev., 1981, 8,229.
11. (a) T. Albers, Doctoral Thesis, 2001, Cardiff University, (b) J. Johnstone,
Doctoral Thesis, 2004, Cardiff University.
12. E. P. Kyba, M C. Keiby, S. P. Rines, Organometallics 1986, 5, 1189.
13. E. L. Muetterties, J. R Bleeke, E. J. Wucherer, Chem. Rev. 1982. 82. 499-525
14. (a) M. J. Atherton, J. H. Hollow, E. G. Hope, A. Karacar, D. R. Russell, G. C.
Saunders, J. Chem. Soc. Chem. Common. 1995,191-192. (b) M. J. Atherton, K. S.
Coleman, J. Fawcett, J. H. Holloway, E. G. Hope, A. Karacar, L. A. Peck, G. C.
Saunders, J. Chem. Soc. Dalton Trans. 1995,24,4029-38.
15. R P. Hughes, D. C. Linder, Organometallics 1996, 15, 5678-5686.
16. R M. Bellabarba, G. C. Saunders, S. Scott, Inorg. Chem. Common. 2002, 5,15-
18.
17. R M. Bellabarba, M. Nieuwenhuyzen, G. C. Saunders, Organometallics 2002,
21(26), 5726-5737.
18. R M Bellabarba, M. Nieuwenhuyzen, G. C. Saunders, Organometallics 2003,
22(9), 1802-1810.
19. O. V. Gusev, S. Sergeev, I. M. Saez, P. M. Maitlis, Organometallics, 1994, 13,
2059-2065.
89
Chapter 3 Cvclobutadienvlcobalt Template Complexes
20. D. S. Trifan, L. Nicholas, J. Am. Chem. Soc. 1957, 79, 2746-9.
21. A. Sdriguchi, T. Matsuo, H. Watanabe, J. Am. Chem. Soc. 2000,122, 5652-565322. R. Gleiter, D. Kratz, Angew. Chem. Ed. Int. 1990, 29, 276-279.
23. Sheldrick, G. M. SHELXTL97, Program for the refinement of Crystal Structures,
University of Goettingen, Germany, 1997.
24. L. D. Freedman, C. O. Doak, J. Am. Chem. Soc 1952, 74, 3414.
90
Chapter 4
Tricarbonyl manganese Template
Complexes
Chapter 4 Triearbonvltnanpinese Template complexes
Abstract
Sequential substitution reactions of [(CO)3Mn(CH3CN)3]+ (4.1) with 1,2-
bis(di-2-fluorophenyl)phosphinobenzene (dfppb) and phenylphosphine in
CH2CI2 lead to the formation of complexes [(CO)3Mn(dfppb)CH3CN]+
(4.2) and [(CO)3Mn(dfppb)(PH2Pli)]+ (4.3) respectively. Treatment of 4.3
with 2 molar equivalents of KOBu1 in THF yields the 9-membered P3
macrocycle complex, [(CO)3Mn{ 1,4-bis(2-fluorophenyl)-7-phenyl-
[b,e,h]tribenzo-1,4,7-triphosphacyclononane}]+ (4.4). UV irradiation of
4.4 in benzonitrile at high temperature followed by work-up in air caused
the formation of a (syn, anti) macrocycle with two phosphorus donors
being oxidized to phosphine oxides and the catalytic transformation of
benzonitrile to benzamide resulting in the formation of complex [(1,4-
dioxo-1,4-bis(2-fluorophenyl)-7-phenyl-[b,e,h]tribenzo-l ,4,7-
triphosphacyclononane}2Mn(C6H5CONH2)2]2+ (4.5) with a Mn(D) center
coordinated by 6 O atoms from two macrocycle ligands and two
benzamide ligands.
91
Chapter 4 Tricarbonylmatiganese Template complexes
4.1 Introduction
To date, a wide range of 1,4,7-triphosphacyclononane macrocyclic
complexes have been synthesized by kinetic template methods using
CpRFe+ (R = H, alkyl) 11"51 or Cb*Co+ (Cb* = q4-C4Me4) (Chapter 3)
based templates. The methodology, originally developed in our
laboratory, is based upon sequentially introducing appropriate phosphine
precursors to the template complex followed by an intramolecular ring
closing reaction. The template-directed method for the synthesis of P3
macrocycles has proved highly efficient, straightforward and also occurs
with good stereocontrol. The demetallation of the product macrocycle
complexes formed by this synthetic methodology remains a significant
problem. All the attempts to liberate the free P 3 macrocycles with ring
systems smaller than 1 2 -members from their iron or cobalt complexes
have been unsuccessful. The difficulty experienced in attempted
liberation of the free ligands is presumably linked to these complexes
being very robust and this feature no doubt depends upon the ring size of
macrocycle {i.e. in our case, the 9-membered ring systems appear to
optimise stability as might be expected). This also demonstrates the good
ligating properties of the P 3 macrocycles with small ring systems and that
ligands of this type will have a rich organometallic and coordination
chemistry. Thus, the study of other template systems remains an
important goal in die search for suitable template systems that may lead
to the facile synthesis of these P 3 macrocyclic ligands, but also allow their
liberation. In this context, we have continued our study of piano-stool
structures containing inert facially capping ligands trans to three labile
ligands, in the search for alternative templates. We have investigated the
substitution and reaction chemistry of these complexes relevant to
92
Chapter 4 Tricarbonylnratnganese Template complexes
phosphorus macrocycle synthesis and we have focused our attention on
cationic manganese carbonyls.
As mentioned previously, in all of the template systems investigated to
date, the liberation of the free phosphorus macrocycles has only been
achieved with the 12-membered, 1,5,9-triphosphacyclododecane and its
tertiary phosphine derivatives from the metal carbonyl templates, fac- (CO^Cr or fac-(CO)3Mo, and less efficiently from the T]5-CpFe+ template
fragment.16"91
The comparison offac-(CO)i$A (M= Cr, Mo), Ti5-CpRFe+ and t|4-Cb*Co+
based templates indicates why the metal-tricarbonyls are not suitable for
the formation of the smaller ring sizes achieved with the Fe and Co
templates. Firstly, the spectator ligands in CpRFe+ and Cb*Co+ are more
sterically bulky than those in yao(CO)3Cr and ^frc-(CO)3Mo (CpR in
CpRFe+, Cb* in Cb*Co+, 3 x CO in the carbonyl complexes). Secondly,
the phosphine precursor complexes, based on CpRFe+ or Cb*Co+, have
relatively shorter M-P bond lengths than do the corresponding complexes
based on fac-ipO^Cx oryoc-(CO)3Mo due to the smaller radius of Fe and
Co in comparison to both Cr or Mo in these complexes. These two factors
cause the spectator ligands in CpRFe or Cb*Co to push the phosphine
precursors into closer proximity and hence facilitate the formation of the
smaller 9-membered P3 macrocycles.
The ^ac-(CO)3Mn(CH3CN)3+ fragment is isoelectronic and structually
analogous to yao(CO)3Cr and c-(C O )3Mo. The cationic Mn(I) centre
however has an intrinsically smaller radius than does the neutral metal in
the group 6 metal carbonyls (1.27 A for Mn, 1.28 A for Cr and 1.39 A for
Mot101); this should help facilitate the closure of smaller ring systems. In
addition, the (CO)3M (M = Cr and Mo) templates are the only known for
93
Chapter 4 Tiicarbonylman^anese Template complexes
which liberation of a P3 macrocycle has been reported and the (CO)3Mn+
system may offer this advantage over the CpRFe+ and Cb*Co+ systems.
An alternative manganese complex, (n5-cyclopentadienyl)Mn(CO)3, has
previously been investigated in our group for the template-directed
synthesis of P3 macrocycles. The removal of all the three CO ligands by
phosphine precursors however, proved to be difficult and attempts to
prepare P 3 macrocycles using this template have failed.[11]
Refluxing (CO)sMnBr in acetonitrile gives the electrophilic fragment,
[(CO)3Mn(CH3CN)3]+ which can be isolated as its PF6- salt in high
yield.[12J The acetonitrile ligands in this species are labile and can be
readily substituted. This feature allows us to develop its chemistry as a
template for the synthesis of P 3 macrocycles with small ring systems.
94
Chapter 4 Tricarbonvlmanganese Template complexes
4.2 Synthesis of a tribenzannulated 9-mem be red triphospha-
macrocycles based on/oe-tricarbonylmanganese templates
Scheme 1 shows the method for the synthesis of fac- tricarbonylmanganese(I)-1,4-bis(2-fluorophenyl)-7 -phenyl-[b,e,h]tri-
benzo-1,4,7-triphosphacyelononane by using the yac-Mn(CO)3+ fragment
as a template.
Scheme 1 Synthesis of a manganese complex of tribenzannulated 1,4,7-
triphosphacyclo-nonane
The strategy is the same as that developed for the synthesis of Fe and Co
analogues which has been addressed in Chapter 1 and Chapter 3. The
weakly coordinated acetonitrile ligands in [(CO)3Mn(CH3CN)3]+(4 .1) act
as leaving groups to be sequentially displaced by l,2 -bis(di-2 -
fluorophenyl)phosphinobenzene (dfppb) and phenylphosphine ligands in
stoichiometric reactions and affording the bisphosphine-monoacetonitrile
complex, 4.2 and bisphosphine-monophosphine complex, 4.3
H2PPh/CH2Cl2
L = CH3CN
4.1 4.2
KOBu*
CO
4.3 4.4
Chapter 4 Tricarbonylmanpanese Template complexes
respectively. The macrocycle complex, 4.4 was subsequently produced
via an intramolecular P-C bond ring-closure reaction from 4.3.
Treatment of [(CO)3Mn(CH3CN)3]PF6 (4.1PF6) with a molar equivalent
of dfppb in CH2CI2 produces the orange complex 4.2,
[(CO)3Mn(dfppb)CH3CN]+, which was isolated as its PF6‘ salt. No
obvious colour change was observed during the reaction. The substitution
of C H 3 C N ligands from 4.1 by dfppb is relatively slow and takes several
hours to reach completion. It was found that removal of the volatile
components in vacuo followed by addition of fresh solvent twice during
the reaction facilitated the substitution, resulting in shorter reaction times.
To prevent any structural isomerizations of 4.2 at elevated temperature,
reactions were performed at room temperature.
The process was monitored by 31P NMR spectroscopy. 31P{1H} NMR
spectroscopy shows the disappearance of a resonance around 5 -34 ppm
due to free dfppb and the growth of a new resonance at 5 72.9 ppm,
which indicates coordinated dfppb (Figure 1). The starting material
(4.1*PF6) was commonly contaminated with other manganese carbonyl
species such as (CO)sMn(CH3CN)2Br, and in which case another1resonance was observed at 8 59.0 ppm in the P{ H} NMR spectrum of
the product. The product was purified by washing with Et2 0 /petrol
leaving the less soluble PF6’ salt of 4.2. The signal due to the CO
resonance in the 13C{1H} NMR spectrum was broadened by the influence
of the quadrupolar 55Mn nucleus[13] and was observed at 8 215 ppm.
96
Chapter 4 Tricarbonylmanganese Template complexes
Figure 1 31P{'H} NMR spectrum of 4.2 in CD2C12 at 20 °C.
The configuration of 4.2 about the manganese centre can be assigned as
fac,cis- in which the dfppb ligand is cis to the CH3CN ligand and the CO
ligands are in a facial arrangement. This stereochemistry is confirmed by
the P NMR and IR spectra. Thus, a single resonance due to coordinated
dfppb in 4.2 is observed in the 31P{1H} NMR spectrum suggesting that
only one isomer is present in solution. For (C0 3)Mn(L2)X (L2, chelated
phosphines; X, monodentate ligand) complexes, the CO vibrational
frequencies and relative intensities in the IR spectrum are very useful for
distinguishing between isomeric metal carbonyl species based on local
symmetry of the CO groups.*141 Complex 4.2 has the expected strong U(co)
absorption bands with approximately equal intensities at 203 l(s), 1961(s)
and 1915(s) cm'1 which supports its fac,cis- stereochemistry.*14'161 The
absorption band at 2344 cm"1 was assigned to u (Cn ) of the CH3CN ligand.
Further displacement of the acetonitrile ligand in 4.2 was achieved by
reaction in CH2Cb with a molar equivalent of phenylphosphine. The
substitution reaction progressed quite rapidly and was monitored by
97
Chapter 4 Tricaibonylmangamese Template complexes
31P{1H} spectroscopy. Complex 4.3, [(CO)3Mn(dfppb)(PH2Ph)]+,
crystallized as its pale yellow PF6’ salt in quantitative yield. The IR
spectrum of 4.3 shows three CO vibrations at 2037(s), 1974(s), 1964(s)
cm'1 which indicates a fac,cis- geometry as mentioned above; the
disappearance of U(cn> and growth of a new band at 2365 cm'1 which
assigned to the PH stretching vibration, confirms die replacement of the
C H 3C N ligand by a phenylphosphine ligand. The slight increase in D(co)
upon replacement o f C H 3 C N by H2PPh is consistent with H2PPh being a
slightly better rc-acid (or poorer a-base) than C H 3C N . The NMR
spectrum of 4.3 shows a broad doublet of PH protons (5 = 4.31 ppm, lJu? = 360 Hz). The resonances attributed to CO were observed at 8 265 ppm
1 -7 |
in C{ H} NMR spectrum although these were broadened due to the
effect of the 55Mn quadrupolar nucleus as well as coupling to P and were
difficult to observe (weak). Again, the signals of coordinated dfjppb and• 1 ____
phenylphosphine in the P{ H} NMR spectrum appear as broad
multiplets at 8 72.4 ppm and 5 -17.6 ppm respectively (Figure 2). The
resonances are complicated due to the influence of the 55Mn quadrupolar
nucleus superimposed on P-P and P-F coupling.
Figure 2 31P{1H} NMR spectrum of 4.3 in CD2C12 at 20 °C.
— 1— <— ■— >— 1— ■— ■— ■— ■— ■— n
1IM •
The geometry of 4.3 was further confirmed by its crystal structure which
was determined by X-ray diffraction (Figure 3). The Mn centre of 4.3
possesses a distorted arc-octahedral coordination environment with facial
98
Chapter 4 Tricarbonvlmanpane.se Template complexes
CO ligands occupying three sites, the dppb ligand chelates in the pseudo- equatorial plane, and the phenylphosphine resides in the remaining
position. The structure of 4.3 bears mirror symmetry, the mirror plane
being normal to the plane of the phenyl group of phenylphosphine; it
passes through the P-Qpso bond, one CO and the Mn centre. The dfppb
ligand has relatively longer M-P bond lengths {2.332(1) A} than does the
phenylphosphine ligand {2.322(1) A}, which may simply be due to the
greatly steric demand of the dfppb ligand than that of the
phenylphosphine ligand although electronic effects may be significant.
The electron-withdrawing ability of the fluorine atoms may help to
decrease the o-basicity of dfppb. The Mn-C bond length trans to the
phenylphosphine ligand {1.840(4) A} is longer than those trans to the
dfppb ligand {1.821(3) A} suggesting that H2PPh may be a better 7t-acid
than dfppb whereas the Mn-C bond length trans to X is shorter than
those trans to the P atoms of L2 in the complexes (CO)3Mn(L2)X (L2,
chelate phosphines; X, monodentate ligand such as Cl, Br, NCO or
PPh3 ).[14,17191 The P-Mn-P angles of 83.61(4)° for the dfppb chelate bite
in 4.3 is smaller than 90° and is similar to other related 5-membered
chelates in (CO^MnfCfiH^PPh^-tf}Cl {81.98(6)°}, (CO)3Mn{C6H4-
(PH2)2-o}C1 {83.06(6)°} and (CO)3Mn(dppe)Br {84.14(10)°}.[171
99
Chapter 4 Tricarbonvlmanganese Template complexes
Figure 3 ORTEP plot (30% probability level) of 4.3 (H atoms on phenyl
group omitted).
C12
C18
C4 C5C2
f H1AC7
D1 c t
Selected bond distances (A): Cl-M nl 1.840(4) C2 -Mnl 1.821(3) Pl-M nl 2.3221(12) P2-Mnl 2.3328(8) Selected bond angles (°): C2-Mnl-C2#l 92.19(17) C2-Mnl-Cl 86.55(13) C2-M nl-Pl 86.19(9) C2-Mnl-P2 92.10(9) C l-M nl-Pl 169.52(14) Cl-M nl-P2 92.93(11) C2#l-Mnl-P2 175.64(9) Pl-M nl-P2 94.88(3) P2#l-M nl-P2 83.61(4) (#1 x, -y, z)
100
Chapter 4 Tricarbonvlman^anese Template complexes
The cyclization reaction to produce the macrocycle complex,
[(CO)3Mn{ 1,4-bis(2-fluorophenyl)-7-phenyl-[b,e,h]tribenzo-1,4,7-
triphosphacyclononane}]+ (4.4), was carried out by addition of 2 molar
equivalents o f KOBu1 to a solution of 4.3PF6 in anhydrous THF,
resulting in a rapid change of colour from yellow to red within a minute.
The solution subsequently changed colour to pale yellow and the less
soluble complex, 4.4*PF6 precipitated out as an off-white powder.
Complex 4.4*PF6 is obtained as an air-stable solid although in solution a
colour change (from colourless to pale yellow) was noticed in the
presence o f air and an unidentified species was observed in the 19F (s, - 8
102 ppm) and 31P{1H} (5 28.4 ppm) NMR spectra. 4 .4 BPh4 was obtained
by anion exchange in order to enhance the solubility of the cation in
common organic solvents. This salt was used for micro-analysis. It is
reasonable to propose that the mechanism of ring closure is similar to that
for the synthesis of the Fe and Co tribenzannulated l,4 ,7 -[9 ]ane-P3
macrocyclic analogues. The process involves deprotonation of
phenylphosphine by base to produce a coordinated phosphide
nucleophile, followed by intramolecular nucleophilic aromatic
substitution of the ortho fluorine atom in the fluorophenyl group within
the coordination sphere of manganese. Although the Mn-P bond lenghs in
complex 4.3 are expected to be relatively longer than the Fe-P and Co-P
bond lengths in analogous CpRFe and Cb*Co complexes, the less-
restricted rotation of the -C6H4F group around the P-Qpso bond (due to the
lower bulk of the CO and perhaps an electrostatic attraction between the
F atom and the phosphide lone pair) allows the nucleophile and
electrophile to orientate in close proximity to favour the cyclisation
reaction. The yield of 4.4 is sensitive to the amount of KOBu1 used. With
excess KOBu1, the dppb ligand in 4.3 may be displaced from the Mn
101
Ft!Chapter 4 Tricarbonvlmanganese Template complexes
icentre. Under these conditions, the free ligand was observed by P{ H}
NMR spectroscopy.
Complex 4.4 was characterized by spectroscopic and analytical methods.
The IR spectrum shows three bands due to i)(co) {2030(s), 1974(s,),
1962(s) cm 1} indicating the fac- configuration. The P3 macrocycle ligand
should have an excellent ligating ability and be a stronger o-donor than
the combination of dfppb and H2PPh ligands in 4.3. This would cause the
metal centre in the macrocycle complex to be more electron rich than that
in 4.3. This renders a slight reduction in U(co) in 4.4 in comparison to that
in 4.3. The molecular ion (m/z = 728 amu) was observed in the mass
spectrum. The disappearance of resonances due to the coordinated dfppb
and phenylphosphine ligands in 4.3 and die growth of resonances
significantly shifted downfield in the 31P{!H} NMR spectrum (Figure 4),1suggests the formation o f the macrocycle. The resonance in the P{ H}
spectrum ( 8 109 ppm) is complex due to P-P and P-F coupling and
quadrupolar broadening which cause the superimposition of multiplets
due to the two unique sets of P atoms in the macrocycle with an A2B
resonance patten. The single resonance of 4.4 in the 19F{!H} NMR
spectrum ( 8 -96.4 ppm) indicates rapid rotation of fluorophenyl groups
around the P-Cexo bonds and relatively little steric hindrance between the
CO ligands and the fluorophenyl group.
102
Chapter 4 Tricarbonvl manganese Template complexes
Figure 4 31P{1H} NMR spectrum of 4.4 in CD3NO2 at 20 °C.
1
10ft*
The crystal structure o f 4.4 is presented in Figure 5 which shows the
tribenzannulated l,4 ,7 -[9 ]ane-P3 ligand binding the Mn centre via three
fused 5-membered chelate rings with three trans (to P) CO ligands
making up the distorted octahedral configuration around the Mn centre.
The Mn-P bond lengths in 4.4 {2.259(1) A, 2.258(1) A, 2.261(1) A} are
significantly shorter than those in the acyclic 4.3 {aver. 2.327(1) A} and
in the { l^ ^ -f^ a n e -P sI^ M ^ C O ^ B r1201 complex {2.327(2) A, 2.243(2) A, 2.331(2) A}. In the example of l ,5 ,9 -[1 2 ]ane-P3Et3, the Mn-P
bonds are associated with 6 -membered chelate rings. The relatively short
Mn-P distances in 4.4 indicate the good ligating capability of the 9-
membered P3 macrocycle. The non-bonded P ...P distances of 3.019(15)
A, 3.024(15) A and 3.030(15) A in 4.4 are comparable to those in the
tribenzannulated macrocycle Co analogues {3.6 in chapter 3, aver.
103
Chapter 4 Tricarfaopylniaiipanese Template complexes
3.000(12) A} but are slightly longer than those in the iron complex,
[CpRFe(l,4 ,7 -[9 ]aneP3-Et3)]+ {aver. 2.969(3) A}, which has an aliphatic
backbone.1211 The slightly longer non-bonded P ...P distances in aromatic
macrocycle complexes (4.4 and 3.6) may be due to the rigidity of the
phenylene backbone functions. The P-Mn-P angles of 83.79(5)°,
83.96(5)° and 84.23(5)° are smaller than those in the Co macrocycle
complex, 3.6 (88.10(5)°, 87.74(4)°, 88.35(4)°}, which is expected due to
Mn(I) having a largo* radius than Co(III) hence a longer Mn-P bond
length than Co-P bond length. The P-Mn-P angles in the more restricted
5-membered rings in 4.4 are also smaller than those in the 6 -membered
chelate rings in ( l- 5 ,9 -[1 2 ]ane-P3Et3}Mn(CO)2Br (aver. 93.81(7)°}.[20]
The average value of 119.2(1)° for the Mn-P-Ce®, (exocyclic phenyl or
fluophenyl group) angles indicates steric repulsion between the terminal
phenyl or fluorophenyl groups and the CO spectator ligands. The Mn-P-
Cexo angles are however considerably smaller than those in the Co
analogue (3.6, aver. 125.60(17)°} indicating that the corresponding
repulsion in 4.4 is weaker than that in 3.6 which bears a much bulkier
tetramethylcyclobutadienyl ligand.
104
Chapter 4 Tricarbonvlmanganese Template complexes
Figure 5 ORTEP plot (30% probability level) of 4.4 (H atoms omitted)C12 C13
C31
(b)Selected bond distances (A): Mnl-C3 1.825(6) M nl-C2 1.832(6) M nl- C1 1.837(6) Mnl-P2 2.2583(15) M nl-Pl 2.2590(15) Mnl-P3 2.2616(15) Selected bond angles (°): C(3>Mn(l>C(2) 93.6(2) C(3)-M n(l)-C(l) 92.5(2) C(2>M n(l>C(l) 92.2(2) C(3>Mn(l)-P(2) 90.95(16) C(2> Mn(l>P(2) 174.09(17) C(l>M n(l)-P(2) 91.32(17) C(3>M n(l)-P(l) 91.18(17) C(2)-M n(l>P(l) 91.97(18) C(l)-M n(l)-P(l) 174.25(18) P(2)- M n(l>P(l) 84.21(5) C(3)-Mn(l)-P(3) 173.15(17) C(2)-Mn(l)-P(3) 91.34(17) C(l)-M n(l>P(3) 91.99(17) P(2>Mn(l>P(3) 83.78(5) P (l> Mn(l>P(3) 83.95(5) C(4>P(l)-M n(l) 119.33(17) C(10>P(l>M n(l) 110.15(18) C(39>P(l)-M n(l) 110.09(19) C(16)-P(2)-Mn(l) 119.10(18) C(15>P(2>Mn(l) 109.89(18) C(22>P(2)-Mn(l) 110.30(18) C(28)-P(3)- M n(l) 119.22(17) C(27)-P(3)-Mn(l) 110.15(18) C(34)-P(3)-Mn(l) 109.73(19)
105
Chapter 4 Tricarbonyl manganese Template complexes
4 3 Preliminary reaction studies towards macrocyclic ligand
liberation
A series of reactions involving metal oxidation, CO abstraction and
fluorine atom substitution of 4.4 have been surveyed in order to explore
macrocyclic ligand liberation methods. Preliminary results show that the
(CO)3Mn+ fragment is not as stable as the CpRFe+ and Cb*Co+ fragments.
Complex 4.4 is susceptible to nucleophilic attack and strong bases such as
KOBu1. The CO ligand can be removed by CO abstraction reagents such
as Me3NO.p2] Variations of the macrocycle may be available by
nucleophilic aromatic substitution at die ortho position of the exo fluorophenyl group by hydride reagents and other nucleophiles. The
oxidation o f the d6 Mn(I) centre to higher oxidation states is also possible
under mild conditions.1231 All these present an opportunity for the
liberation o f the P3 macrocycle and for future exploration of its
coordination and/or catalytic chemistry. Interestingly, UV irradiation of
4.4 in benzonitrile at high temperature for 3 hours, followed by work-up
in air, gives rise to a new complex, [(l,4 -dioxo-l,4 -bis(2 -FC6H4)-7 -C6H5-
[b,eJi]tribenzo-l,4 ,7 -triphosphacyclonoiiane)2Mn(C6H5CONH2)2]2+ (4.5),
which was isolated as a minor product (Scheme 2).
Scheme 2 liberation o f P3 macrocycle by photolysis and oxidation
h 2n
Chapter 4 Tricarbonvlmanganese Template complexes
Complex 4.5 was crystallized as its SbF6' salt and characterized by X-ray
diffraction (Figure 6 ). The cation bears an octahedral Mn(II) centre
located on a crystallographic inversion centre with a MnC>6 environment.
The CO ligands in 4.4 have clearly been removed. Two P atoms of each
triphosphamacrocycle in 4.4 have been oxidised to phosphine oxides
which are coordinated to the Mn atom in a cis arrangement through the O
donors. It is reasonable to assume that the two P atoms which bear
electron-withdrawing fluorophenyl groups were oxidised although the
positions o f these two F atoms could not be unequivocally located as they
were disordered in the crystal structure and refined in partial occupancies
(0.8, 0.7, 0.5) attached to all three phenyl substitutents. All the P=0
oxygens lie in the equatorial plane. The third P atom is uncoordinated and
is inverted relative to the others two, presumably due to die high
temperature during the reaction, although oxidation of the other two
phosphines may cause this inversion depending upon the mechanism of
oxidation. The resultant conformation at the three P atoms is syn, anti. Surprisingly, two benzamide ligands are also coordinated to the Mn
centre via the O donors; they occupy the axial positions.
It is not clear what mechanism operates during this reaction, or what kind
of intermediates are involved during the reaction. The reaction which
yielded 4.5 however is important in that it clearly indicates that the P3
macrocycle ligand has been liberated during the reaction. In this study, it
appears that these manganese complexes show catalytic activity for the
hydration of benzonitrile to benzamide.1241 A trace of water present in the
system is presumably responsible for the promoting this transformation.
Chapter 4 Tricarbonvlmanganese Template complexes
Figure 6 ORTEP plot (30% probability level) of 4.5 (H atoms on phenyl
group omitted).
(b)
Selected bond distances (A): M nl-O l 2.106(2) M nl-02 2.246(2) M nl- 03 2.154(3) 03-C71 1.250(4) N1-C71 1.322(5) Selected bond angles (°): 01#1-M nl-01 180.0(3) 01#l-M nl-03 92.38(10) O l-M nl-03 87.62(10) 01-M nl-03#1 92.38(10) 03-M nl-03#l 180.0(2) 01-M nl-02#194.49(9) 03-M nl-02#l 91.62(9) O l-M nl-02 85.51(9) 03-M nl-02 88.38(9) 02#l-M nl-02 180.0(2) (#1 -x, -y, -z)
108
Chapter 4 Tricarbonvlmanganese Template complexes
4.4 Conclusions and ProspectsFor the first time, the carbonyl manganese complex, fac-(CO)$sArL has
been successfully used as a template to synthesize a 9-membered P3
macrocycle, (CO)3Mn-1,4-bis(2-fluorophenyl)-7 -phenyl-[b,e,h]tribenzo-
1,4,7-triphosphacyclononane (4.4). The work is of interest since it may be
easier to labilise the Mn (than the Fe or Co in previously studied
templates) and hence to liberate the macrocycle ligand. Preliminary
experiments show that complex 4.4 undergoes very interesting reactions
towards the liberation of the macrocycle ligand.
For the CpRFe+, Cb*Co+ and (CO)3Mn+ templates, the strategy for the
synthesis of 9-membered P3 macrocycles with tribenzannulated
backbones by an intramolecular dehydrofluorination works well. In
contrast to the Fe system, the Co and Mn systems seem to be less suitable
for the synthesis of other P3 macrocycle derivatives with small ring
systems and aliphatic backbones. The template-directed method is based
on substituting labile acetonitrile ligands with a suitable diphosphine
followed by introduction of a complementary monophosphine to the
metal centre in the first stage. The limitations of this method arise from
the different ability of appropriate phosphines to form reasonably stable
complexes with the template metal, and also to substitute acetonitrile.
Thus to extend the chemistry of the Cb*Co+ and^ac-(CO)3Mn+ fragments
as more versatile templates for the synthesis other P3 macocycle
derivatives with small ring systems remains a challenging task.
Chapter 4 Tricarbonylmanganese Template complexes
4.5 Experimental Section
For general experimental information see C hapter 2 and 3.
M aterials: [(CO)3Mn(CH3CN)3]PF6[12] and phenylphosphine1251 were
prepared according to literature methods.
X-ray crystallography: Complex 4.3 was crystallised as its PF6‘ salt.
Both o f complex 4.4 and 4.5 were crystallised as SbF6" salts. The two F
atoms of fluorophenyl group in 4.4 are refined and distributed in five
ortho positions of three phenyls with disordered model of approximately
occupancy factor of 0.64:0.22:0.14:0.56:0.44 for FI A, FIB, F1C, F2A
and F2B respectively. Crystal structure and refinement data are collected
in Appendix C and supplementary CD.
[(CO)3M n { (i^ 6H4F)2PC6H4P (^C 6H4F)2}(CH3CN)]PF6 (4.2 PF6)
To a solution of [(CO>3Mn(CH3CN)3]PF6(4.1PF6)(0.15 g, 0.368 mmol)
in 20ml CH2C12, dfppb (0.19g, 0.368 mmol) was added with stirring.
After 2 hours, the solvent was evaporated in vacuo. Then 20ml CH2C12
was added and stirring continued for another 2 hours. The 31P{1H) NMR
spectrum showed the growth of a resonance due to coordinated dfppb at
6 72.9 ppm. All the volatiles were removed in vacuo, and the solid
residue washed with Et20/petrol and then dissolved in CH2C12, followed
by filtration and concentration. Orange crystals of 4.2 PF6 were obtained
by diffusion of diethyl ether vapour into the solution. Yield: 98%.
31P{'H} (CD2CI2): 72.9(dfppb), -144.1(sep, PF6, = 711 Hz). 19F{'H}
(CD2CI2): -94.5 (s, dfppb), -95.0 (s, dfppb), -73.1(d, PF6). 13C{‘H}
(CD2CI2): 215 (CO, weak and broad), 2.4 (MeCN). 'H (CD2CI2): 7.6 to
6.4 (20H, m, H ^,), 1.15 (3H, s, CH3CN). IR: 2344 (w, CN), 2031 (s,
CO), 1961 (s, CO), 1915 (s, CO)
110
Chapter 4 Tricarixmylmanganese Template complexes
[(CO)3Mii{(a-C6H4F)2PC6H4P(^C6H4F)2}(PH2Ph)]PF6 (4.3 PF6)
To a solution of 4.2*PF6 (0.20g, 0.22mmol) in 20ml CH2CI2,
phenylphosphine (0.03g, 0.27mmol) in toluene was added. After stirring
for 15 minutes, P{ H} NMR spectrum showed the growth of a broad
resonance of coordinated phenylphosphine at 5-17.6 ppm. The resonance
for coordinated dfppb became very broad at 5 72.4 ppm. Then the
reaction mixture was filtered. The solvent was removed in vacuo and the
residue was triturated with diethyl ether to remove excess amount of
phenylphosphine. The yellow powder was then dissolved in CH2CI2 and
concentrated. Yellow crystals of 4.3 PF6 were obtained by diffusion of
diethyl ether vapour into the solution. Yield: 98%. 31P{1H} (CD2CI2):
72.4 (dfppb), -17.6 (dfppb), -143.3 (sep, PF6, = 710 Hz). l9F{lH} (CD2CI2): -93.2 (s, dfppb), -94.4 (s, dfppb), -73.2(d, PF6). 13C{!H}
(CD2CI2): 265 (CO, very weak). *H (CD2CI2): 7.4 to 6.7 (m, 25H, H ^ ),
4.31 (d, 2H, PH, lJm = 360 Hz). IR: 2365 (s, PH), 2037 (s, CO), 1974 (s,
CO), 1964 (s, CO). M.S.(APCI): 767.9 (M*)
(CO)3M n-l,4 -bis(2 -fluorophenyl)-7 -phenyl-[b,e,h]tribenzo-l,4 ,7 -
triphosphacyclononane hexafluorophosphate (4.4PF6)
To a solution of 4.3 PF6 (0.27 g, 0.29 mmol) in 30ml THF was added
KOBu4 (65 mg, 0.58mmol). The colour of the solution changed over
approximately 1 minute from yellow to red then to very pale yellow with
the formation of a white precipitate. After filtering, the white precipitate
was dissolved in CH2C12 and a 31P{!H} NMR spectrum showed a broad
complex multiplet significantly shifted downfield at 8 109 ppm. 4 .4 BPI14
was collected by anion exchange from its PF6" salt in THF to enhance the
solubility. Single crystals for X-ray analysis ware obtained as its SbF6_
salt. Yield: more than 80%. ( C D 3 N O 2 ) : 109 (m, broad), -144.7
(sep, ‘Jpj = 712 Hz). 13C{'H} (CD2CI2): 251 (CO, very weak). 19F
Chapter 4 Tricarbonvlmanganese Template complexes
(CD2C12): -96.4 (s, fluorophenyl), -73.4 (d, PF6). *H {CD2C12}: 7.6 to 7.1
(m, Haryi). IR (KBr): 2030 (s, CO), 1974 (s, CO), 1%2 (s, CO).
M.S.(APCI): 728.0 (M*). Anal. Calcd for C63H45BiF2MnI0 3P3 (4.4 BPh4,
M= 1046.71 gmol1): C, 72.29; H, 4.33. Found: C, 72.08; H, 4.23.
[(CO)3M n{l,4 -dioxo-l,4 -bis(2 -fluorophenyl)-7 -phenyl-[b,e,h]tri-
benzo-1,4,7-triphosphacyclononane} ] SbF6 (4.5*SbF6)
A solution of 4.4SbF6 (40 mg, 0.04 mmol) in 20 ml benzonitrile was
irradiated with UV light at 180 °C for 3 hour. The colour of the solution
changed from colourless to red. Mass and 31P{1H} NMR spectra
indicated that 4.4 still existed as the major component in die solution. The
solvent was removed in vacuo and the residue was dissolved in CH3CN.
Colourless crystals of 4.5*SbF6 were obtained by evaporation of the
solvent slowly in open air.
112
Chapter 4 Tricarbonvlmanganese Template complexes
References
1. P. G. Edwards; M. L. Whatton; R. Haigh, Organometallics 2000, 19(14), 2652-
2654.
2. P. G. Edwards; P. D. Newman; D. E. Hibbs, Angew. Chem. Int. Ed. 2000, 39(15),
2722-2724.
3. R. Haigh, Doctoral Thesis, 2002, Cardiff University
4. T. Albers, Doctoral Thesis, 2001, Cardiff University
5. J. Johnstone, Doctoral Thesis, 2004, Cardiff University
6. S. J. Coles; P. G. Edwards; J. S. Fleming; Hursthouse, B. Michael; S. S.
Liyanage, Chem. Commun. 1996, (3), 293-4.
7. P. G. Edwards; J. S. Fleming; S. S. Liyanage, Inorg. Chem. 1996, 35(16), 4563-
4568.
8. P. G. Edwards; J. S. Fleming; S. S. Liyanage, Dalton Trans., 1997, (2), 193-197.
9. A. J. Price; P. G. Edwards, Chem. Commun. 2000 (11), 899-900.
10. J. G. Speight, Lange *s Handbook of Chemistry 16th Ed. McGraw-Hill, New York
2005.
11. R. Haigh, Doctoral Tthesis 2002, Cardiff University
12. R. H. Reimann; E. Singleton, J. Organomet. Chem., 59, C24-C26,1973
13. D. Rentsch; R. Hany; W. von Philipsbom, Magnetic Resourance in Chemistry, vol. 35, 832-838,1997
14. M. A. Beckett et al, J. Organomet. Chem., 688,174-180, 2003
15. G. A. Carriedo, V. Riera, J. Organomet. Chem., 205, 371-379,1981
16. R. J. Angditi; F. Basolo; A. J. Poe, J. Am. Chem. Soc. 85(15), 2215-19,1963.
17. S. J. A. Pope, G. Reid, J. Chem. Soc., Dalton Trans., 1615-1621,1999
18. S. K. Mandal; D. M. Ho; M. Orchin, Polyhedron, 11(16), 2055-2063,1992
19. T. M Becker; J. A. Krause-Bauer; C. L. Homrighausen; M. Orchin, Polyhedron, 18, 2563-2571,1999.
20. R. J. Baker, P. G. Edwards; J. Gracia-Mora; F. Ingold; K. M. A. Malik, Dalton Trans. 2002, (21), 3985-3992.
21. P. G. Edwards; P. D. Newman; K. M A. Malik, Angew. Chem. Int. Ed. 2000,
39(16), 2922-2924.
22. M. O. Albers; N. J. Coville, Coord Chem. Rev., 53, 227-259,1984.
113
Chapter 4 Tricarbonvlmanganese Template complexes
23. R. H. Reimann; E. Singleton, J. Chem. Soc, Dalton Trans, 808-813,1974
24. Wai Kit Fung; X. Huang; M. L. Man; S. M. Ng; M Y. Hung; Z. Lin; C. Po. Lau,
J. Am. Chem. Soc. 2003, 125(38), 11539-11544.
25. L. D. Freedman; C. O. Doak, J. Am. Chem. Soc 1952, 74, 3414.
114
Appendix ACrystal data and structure refinement for complexes in Chapter 2Complex 2.1-2BF4 2 .2 2 BF4 CH3CN 2.3-2BF4 2.4-2BF4CH3CN 2.5-2BF4Empirical formula C22H3 lN9B2F8Fel C36H34N10B2F8Fel C22H3 lN9B2F8Col C36H34N10B2F8Co 1 C32H54N2B2F8P4Fe 1M, g mol1 651.03 836.20 654.11 839.28 820.12Crystal system Trigonal Monoclinic Trigonal Monoclinic MonochnicSpace group P-3 P2(l)/n P-3 P2(l)/n P2(l)/na, A 11.3020(7) 12.3614(2) 11.3359(5) 12.3345(2) 10.3691(3)ft, A 11.3020(7) 16.4253(3) 11.3359(5) 16.3676(3) 11.8290(3)c, A 13.3820(6) 19.3493(4) 13.2939(6) 19.3388(4) 16.2524(4)A ° 92.5822(10) 92.6709(6) 92.2681(16)Z 2 4 2 4 2K,A3 1480.34 3924.69(13) 1479.43(11) 3899.99(12) 1991.89(9)Aaicd, g cm’3 1.427 1.415 1.435 1.429 1.367^mm'1 0.586 0.463 0.656 0.520 0.602Crystal size (mm3) 0.23x0.19x0.17 0.20x0.18x0.16 0.12x0.10x0.08 0.22x0.18x0.08 0.26x0.22x0.18Reflections collected 6874 31085 6424 61192 18584Unique reflections (Rini) 1931(0.0559) 7718(0.0809) 2255(0.0520) 7619(0.1291) 4542(0.0676)Max./min.transmission 0.9273/0.9018 0.9295/0.9130 0.9494/0.9254 0.9596/0.8943 0.8994/0.8592Goodness-of-fit on F2 1.094 1.053 1.080 1.029 0.997R values [/ < 2o(/o)] Rl = 0.0716
wR2 = 0.2048Rl = 0.0574 wR2 = 0.1336
R l= 0.0752 wR2 = 0.2329
Rl = 0.0495 wR2 = 0.1130
Rl = 0.0608 wR2 = 0.1738
R values, all data Rl = 0.0923 wR2 = 0.2231
Rl= 0.0812 hR2 = 0.1462
Rl =0.1098 wR2 = 0.2515
Rl = 0.0668 wR2 = 0.1225
Rl = 0.0866 wR2 = 0.1838
'"Atomic coordinates, equivalent isotropic displacement parameters, bond lengths and bond angles for all the complexes are collected in the supplementary CD.
115
Appendix BCrystal data and structure refinement for complexes in Chapter 3Complex 3.2PF6 3.3PF6 3.4PF6 3.5PF6 3.6PF6CH3CN 3.7-SbF6-0.5CH3CNEmpirical formula C36H39Col
F6N1P3C40H43ColF6P4
C40H35ColF10N1P3
C44H39ColF10P4
C46H40ColF8N1P4
C45H36.5 Col F6N0.5P3Sbl
M, g mol'1 751.52 820.55 871.53 940.56 941.60 971.88Crystal system Monoclinic Monoclinic Orthorhombic Triclinic Monoclinic MonochnicSpace group P21/c P21/c P cab P-l P21/c C 2/ca, A 11.3864(3) 13.3220(2) 11.3359(5) 10.9766(2) 14.5081(3) 29.757(5)b, A 16.8174(4) 19.5100(4) 11.3359(5) 13.7217(3) 14.9988(3) 14.137(5)c, A 18.6870(4) 14.5130(3) 13.2939(6) 13.8448(3) 19.7361(4) 19.573(5)a ,0 15.9700(2) 84.052(1)
104.2732(10) 96.3780(12) 24.4817(3) 75.647(1) 100.0900(10) 96.818(5)y,° 19.5063(3) 79.706(1)z 4 4 8 2 4 8f,A3 3467.91(14) 3748.76(12) 7626.43(18) 1983.81(7) 4228.23(15) 8176(4)Adcd, g cm'3 1.439 1.454 1.518 1.575 1.479 1.579^,mm'1 0.693 0.688 0.656 0.675 0.628 1.245Crystal size (mm3) 0.40,0.35,0.30 0.35,0.25,0.23 0.35,0.20,0.15 0.35,0.30,0.15 0.20,0.20,0.20 0.25,0.10,0.10Reflections collected 30795 42415 94367 29619 40958 62879Unique reflections (Rind 7921(0.0930) 8574(0.1104) 8722(0.1228) 9034(0.0810) 9481(0.0991) 9339(0.1234)Max./min. transmission 0.962/0.686 0.936/0.829 0.9080/0.7980 0.9055/0.7980 0.998/0.843 0.8856/0.7461Goodness-of-fit on F2 1.062 1.032 1.041 1.050 1.021 1.031R values [7 < 2o(/o)] Rl = 0.0560
wR2 = 0.1229fll = 0.0479 wR2 = 0.1064
Rl = 0.0463 wR2 = 0.0961
Rl = 0.0467 wR2 = 0.0972
Rl= 0.0697 wR2 = 0.1477
Rl = 0.0820 wR2 = 0.1940
R values, All data Rl = 0.0886 wR2 = 0.1359
Rl = 0.0839 wR2 = 0.1210
*1= 0.1087 wR2 = 0.1166
Rl = 0.0794 wR2 = 0.1093
jR1 = 0.1242 wR2 = 0.1695
Rl = 0.1149 wR2 = 0.2163
♦Atomic coordinates, equivalent isotropic displacement parameters, bond lengths and bond angles for all the complexes are collected in the supplementary CD.
116
Appendix CCrystal data and structure refinement for complexes in
C hapter 4Complex 43*PF6 4.4 SbF6-2CH3CN 4.5-2SbF6*2CH3CNEmpirical C39H27F10Mnl C43H3 lF8MnlN20 C90H70F16Mnlformula 03P4 3P3Sbl N406P6Sb2M, g mol'1 912.43 1045.30 2091.76Crystal system Orthorhombic Triclinic TriclinicSpace group Ccmb P-l P-la , A 18.2370(2) 9.0240(2) 13.1630(2)b, A 21.3400(3) 15.6510(3) 14.2040(3)c , A 19.5450(3) 15.9190(4) 14.8460(3)a,° 75.1480(10) 111.1180(10)fi° 80.1430(10) 115.5180(10)y> ° 77.2800(10) 97.725(2)z 8 2 1v,A 3 7606.48(18) 2104.11(8) 2194.06(7)A a lc d , g cm'3 1.594 1.650 1.583p, mm*1 0.603 1.134 0.925Crystal size (mm3) 0.32x0.25x0.20 0.25x0.10x0.05 0.23x0.23x0.20Reflections collected 53306 32820 32132Unique reflections (Rmt) 3995(0.0781) 8594(0.1244) 10016(0.0774)MaxVmin. transmission 0.8888/0.8303 0.9455/0.7647 0.8324/0.8107Goodness-of-fit on F2 1.014 1.081 1.034R values [I < 2o(/0)] R\ = 0.0447 R\ = 0.0605 Rl = 0.0542
wR2 = 0.1088 wR2 = 0.1482 wR2 = 0.1036R values, all data R\ — 0.0550 Rl = 0.0901 Rl = 0.1044
wR2 = 0.1145 wR2 = 0.1629 wR2 = 0.1204♦Atomic coordinates, equivalent isotropic displacement parameters, bond lengths and bond angles for all the complexes are collected in the supplementary CD.
117
Acknowledgements
This thesis was completed under the supervision of Prof. Peter G.
Edwards. I sincerely appreciate his instructions, not only on chemistry but
English and other aspects, and financial support throughout the three
years study.
Grateful thanks are also due to Dr Paul D. Newman for valuable advice
and assistance on daily laboratory work and proof-checking.
I would also like to thank Dr Angelo J. Amoroso for his help in
electrochemistry and kind encouragements in 6 monthly interviews with
Dr Ian A. Fallis. I thank Dr Abdul Malik and Dr Li-ling Ooi for
crystallography, Prof. Andrew Harrison (University of Edinburgh) for
magnetic susceptibility measurements, John for help with the safety
issues, Rob for technical assistance on NMR spectroscopy, Sham and
Robin for the Mass spectra, Gary and Stephen for organizing the
chemicals, Alun for checking the pump and not least, Ricky for making
the glassware.
Thanks to all my colleagues in Lab. 2.91 during the memorable three
years including Andy, Dongmei, Julia, Eli, Becky, Adam, Kate, Chris,
Sultan, Dan, Martie, Abdulla, Sudantha, Rajika, Gayani, Thusith, Peter,
Andrea and Mark.
118